article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
emergent geometry is a key concept in formulating a quantum theory of gravity because in such a theory space(time ) is believed to be not fundamental but emergent .
a concrete realization of this notion has been achieved by the gauge / string duality @xcite , which states an equivalence between gauge theories and string theories .
for instance , the duality between the @xmath10 matrix model and the two - dimensional string theory is known to be a prototype of the gauge / string duality ( see for reviews @xcite ) . in this duality
, the matrix model can be mapped to a free fermion system .
the fermi surface , which is the eigenvalue density of the @xmath10 matrix model , corresponds to one spatial direction of the dual two - dimensional geometry .
another well - known example is the duality between 1/2-bps operators in @xmath1 super yang - mills theory ( sym ) and the bubbling geometries in type iib supergravity @xcite .
the chiral primary operators in @xmath1 sym can also be mapped to the states of free fermions @xcite and its phase space can be identified with the droplets in the gravity dual @xcite .
the same kind of correspondence is shown to exist for the 1/2-bps wilson loop operators in @xcite . in these cases also , the fermi surface corresponds to one spatial direction in the dual ten - dimensional geometry .
other recent examples include the duality between wilson surfaces in six - dimensional @xmath11 theory and m - theory bubbling geometries @xcite .
recently , yet another example of the emergent geometry has been demonstrated by the present authors and t. okada in @xcite .
there , we studied the gauge / gravity duality between the plane wave matrix model ( pwmm ) and the bubbling geometry in type iia supergravity @xcite , which is explained in detail below .
pwmm is a matrix quantum mechanics originally proposed as a formulation of m - theory in the light - cone frame on the pp - wave geometry @xcite .
the theory is the mass deformation of the bfss matrix theory @xcite and has @xmath0 symmetry , which consists of @xmath12 bosonic symmetry . ] and 16 supersymmetries .
pwmm possesses many discrete vacua , called fuzzy sphere , which are labeled by representations of @xmath13 .
one can make the irreducible decomposition of the representations , so that the vacua are specified by irreducible representations and their multiplicities which appear in the decomposition .
for each vacuum of pwmm , the dual gravity solution was constructed by lin and maldacena in type iia supergravity , which is known as the bubbling geometry @xcite .
if @xmath0 symmetric ansatz is assumed , the equations of motion can be reduced to a simple differential equation with boundary conditions characterized by fermionic droplets on a particular one - dimensional line in ten - dimensions . solving this equation
is shown to be equivalent to solving for an electrostatic potential of a three - dimensional axially symmetric electrostatic system with some conducting disks @xcite . in this duality ,
the dimensions of irreducible representations of a fuzzy sphere vacuum correspond to ns5-brane charges and their multiplicities correspond to d2-brane charges . in the associated electrostatic system , they correspond to the positions and the charges of the conducting disks .
see fig.1 for an electrostatic system for a general vacuum of pwmm . in @xcite ,
the duality for pwmm was studied in the case of the vacuum characterized by the direct sum of copies of an irreducible representation .
the gravity dual of this theory was examined in @xcite , where the charge density on the conducting disk in the corresponding electrostatic system was shown to satisfy a fredholm integral equation of the second kind . in pwmm ,
a 1/4-bps sector was considered .
this sector is made of a complex scalar that corresponds to two spatial directions in the dual geometry on which the electrostatic system is defined . by applying the localization technique @xcite
, this sector reduces to a matrix integral @xcite . by evaluating this matrix integral
, it was found that the eigenvalue density of the matrix integral obeys the same integral equation as the charge density in the electrostatic problem .
this fact naturally leads us to identify the eigenvalue density with the charge density .
since the charge density completely determines the gravity solution , it was concluded that one can exactly reconstruct the gravity solution from pwmm .
remarkably , the sector studied in @xcite is interacting and so the coupling constant appears as a parameter in this duality .
this is in sharp contrast to the examples given in the first paragraph , where the relevant sectors in the gauge theories are free and there is no coupling dependence .
the existence of such an extra parameter makes the duality more attractive . actually , the gauge / gravity duality for pwmm admits two interesting limits that lead to the duality for other field theories @xcite .
one is the d2-brane limit ( commutative limit of fuzzy sphere ) , in which pwmm becomes sym on @xmath4 , and the dual geometry asymptotically becomes d2-brane geometry .
the other one is the ns5-brane limit , in which pwmm is considered to become type iia little string theory ( lst ) @xcite on @xmath14 , and the dual geometry asymptotically becomes ns5-brane geometry .
thus , the study of the gauge / gravity duality for pwmm enables us to study that for sym on @xmath4 and even that for lst in a comprehensive manner . in this paper
, we extend the results of @xcite to the case of a general vacuum of pwmm .
we first consider the gravity side .
we solve the electrostatic problem for a general vacuum of pwmm by extending the method developed in @xcite .
we then show that the electrostatic potential of the problem can be described in terms of the charge densities on each conducting disk and the charge densities satisfy a system of integral equations .
next , we study the gauge theory side .
we consider a general vacuum and analyze the 1/4-bps sector as in the previous paper @xcite . in this case
, the 1/4-bps sector can be described by a multi - matrix integral @xcite .
we show that the saddle point equations for the eigenvalue densities of the matrix integral are exactly the same as the system of integral equations for the charge densities in the corresponding electrostatic problem .
thus , as in @xcite , we can identify the eigenvalue densities with the charge densities , which determine the gravity solution .
this result shows that pwmm around a general vacuum can contain its gravity dual geometry as saddle point configurations of eigenvalues .
moreover , we investigate the gauge / gravity duality for other gauge theories with @xmath0 symmetry , @xmath3 sym on @xmath4 and @xmath1 sym on @xmath2 . it was shown in @xcite that the gauge / gravity duality for sym on @xmath4 can be obtained from pwmm by taking the d2-brane limit and that for sym on @xmath2 by taking the t - duality as well as the d2-brane limit . using these relations , we prove the correspondence between eigenvalue densities in these theories and charge densities in the electrostatic systems associated with their gravity duals .
the organization of this paper is as follows . in section 2 ,
we review the dual gravity solutions for gauge theories with @xmath0 symmetry and solve general electrostatic problems associated to these solutions . in section 3 , after reviewing pwmm and the matrix integral , we derive the saddle point equations for the eigenvalue densities .
then , we show that the charge densities can be identified with the eigenvalue densities .
finally , we discuss the cases for sym on @xmath4 and sym on @xmath2 .
in this section , we elaborate the lin - maldacena solution for a general vacuum of the gauge theories with @xmath0 symmetry by analyzing the corresponding electrostatic system .
we will show that the charge densities on each disk , which determine the gravity solution , satisfy a system of integral equations .
first , we review the lin - maldacena solution @xcite , which is the solution with @xmath0 symmetry in type iia supergravity .
the solution is given by @xmath15 where @xmath16 and the dots and primes denote @xmath17 and @xmath18 , respectively .
a remarkable feature of this solution is that it is written in terms of a single function @xmath19 .
the killing spinor equation in the supergravity imposes a condition that @xmath19 satisfies the laplace equation in a three - dimensional axially symmetric electrostatic system , where @xmath20 and @xmath21 represent coordinates for the transverse and the axial directions , respectively .
the regularity of the metric requires that the electrostatic system must consist of some conducting disks with radii tuned such that the charge densities vanish at the edges . in addition , from the positivity of the metric , there must be a certain background potential .
so , the potential @xmath19 consists of these two contributions : @xmath22 the electrostatic system is determined once a theory and its vacuum are specified . the electrostatic system relevant to pwmm consists of an infinite conducting plate at @xmath23 , some finite conducting disks in the region of @xmath24 ( fig.1 ) and the background potential of the form @xmath25 where @xmath26 is a constant .
the electrostatic system relevant to @xmath3 sym on @xmath4 consists of some finite conducting disks in the region @xmath27 ( fig.2 ) and the background potential of the form @xmath28 where @xmath29 is a constant .
the electrostatic system relevant to @xmath1 sym on @xmath2 consists of an infinite number of finite conducting disks arranged periodically along the @xmath21-axis ( fig.3 ) and the background potential .
, which do not contribute to the gravity solution . ] ( left ) and sym on @xmath2 ( right ) .
the conducting disks in the right figure are arranged periodically.,title="fig:",height=264 ] ( left ) and sym on @xmath2 ( right ) .
the conducting disks in the right figure are arranged periodically.,title="fig:",height=264 ] the condition that the charge densities vanish at the edges of the disks relates the radii of the disks and the charges .
so the independent parameters of this solution are the total charges and the @xmath21-coordinates of the disks , as well as @xmath26 in or @xmath29 in . while @xmath26 and @xmath29 are related to the magnitude of the dilaton , the other parameters turn out to be related to the brane charges .
this geometry has an @xmath30 and an @xmath31 at each point on the @xmath20-@xmath21 plane .
one can show that on the @xmath21-axis the radius of @xmath31 becomes zero , while on the finite conducting disks , the radius of @xmath30 becomes zero .
so one can consider a various non - contractible @xmath32 s or @xmath33 s which are given by fibering the @xmath30 or @xmath31 on the path ( on the @xmath20-@xmath21 plane ) ending on different disks or on different segments of the @xmath21-axis , respectively . on each @xmath32 or @xmath33 , one can measure the nsns - flux or the rr - flux .
this relates the parameters of the electrostatic system to the brane charges . as an example , let us consider the electrostatic system for pwmm with @xmath34 finite plates ( see fig .
1 ) . we denote the total charge , the radius and the @xmath21-coordinate of @xmath35-th disk by @xmath36 , @xmath37 and @xmath38 , respectively , where @xmath39 . in this case , there are @xmath34 independent non - contractible @xmath32 s and the same number of @xmath33 s in the geometry . @xmath36 and
@xmath38 are related to the d2-brane charges @xmath40 and the ns5-brane charges @xmath41 as @xmath42 for @xmath43 . here
@xmath44 denotes the position of the infinite plate .
we will show in section [ emergent bubbling geometry ] how the parameters of the electrostatic system are mapped to those of the corresponding gauge theory . in the following ,
we derive a fredholm integral equation which determines @xmath45 in ( [ vbg+tildev ] ) .
we consider the situation shown in fig.1 and denote the solution of @xmath45 in the region @xmath46 by @xmath47 , where @xmath35 runs from @xmath48 to @xmath34 and @xmath49 is assumed .
we first note that a general solution to the laplace equation in an axially symmetric system is given by @xmath50 where @xmath51 is the bessel function of the first kind of order zero and @xmath52 is a positive real number .
so we can write @xmath47 as @xmath53 now we have the boundary condition that @xmath54 as @xmath55 and @xmath56 .
this means @xmath57 we also have some continuation conditions for @xmath58 s at @xmath59 .
first , at @xmath60 , @xmath58 should be equal to @xmath61 .
this is satisfied if @xmath62 second , when @xmath60 and @xmath63 , not only @xmath64 but also @xmath65 should be continuous .
so we have @xmath66 for @xmath63 .
third , when @xmath60 and @xmath67 ( i.e. on the conducting disk ) , the value of @xmath68 should be constant : @xmath69 . in terms of @xmath70 and @xmath71 ,
this is written as @xmath72 in order to solve the conditions ( [ first cond ] ) , ( [ second cond ] ) and ( [ third cond ] ) , we define @xmath73 for @xmath74 . from ( [ bc for cd ] ) and ( [ first cond ] ) , @xmath70 and @xmath71 can be written in terms of @xmath75 as @xmath76 by substituting ( [ def of a ] ) and ( [ cd in terms of a ] ) to ( [ second cond ] ) and ( [ third cond ] ) , we obtain @xmath77 where @xmath78 and @xmath79 are given by @xmath80 as shown in appendix [ dual integral equations ] , the equations ( [ equations for a ] ) can be reduced to the integral equations , ( [ dual eq for h2 ] ) and ( [ dual eq for h1 ] ) , for the functions @xmath81 defined by ( [ def of hs ] ) , @xmath82 and @xmath83 in appendix [ dual integral equations ] corresponds to @xmath34 , @xmath84 and @xmath85 in our problem , respectively . ] . for our problem , it is more convenient to work with the variables @xmath86 then , ( [ dual eq for h1 ] ) is written for @xmath87 as @xmath88f_{t}(u ) \nonumber\\ = \frac{1}{\pi}\left ( \delta_s + \frac{2}{3}v_0 d_s^3 -2v_0 d_s x^2 \right ) , \label{integral eq for fs}\end{aligned}\ ] ] and ( [ dual eq for h2 ] ) shows that @xmath89 is vanishing outside the region @xmath90 $ ] .
here we have defined @xmath91 with negative @xmath92 as @xmath93 and extended the domain to the entire real line .
the function @xmath94 can be interpreted as the charge density on the @xmath35-th conducting disk as follows . for @xmath95 and @xmath96
, we have @xmath97 on the other hand , this is equal to @xmath98 , where @xmath99 is the charge density for the @xmath20 direction .
hence , the total charge on the disk can be computed as @xmath100 these relations show that @xmath101 corresponds to the charge density on the @xmath35-th plate projected onto a diameter direction .
these densities are fully determined by ( [ integral eq for fs ] ) and so is the potential which can be written in terms of @xmath102 as @xmath103f_s(t ) .
\label{pot}\end{aligned}\ ] ] note that @xmath37 and @xmath104 are determined by @xmath105 and .
the electrostatic system associated with the gravity dual of @xmath3 sym on @xmath4 is shown in fig.2 ( left ) .
the case where @xmath106 and @xmath107 was studied in @xcite .
here , we generalize their result .
it was shown in @xcite that the solution for this system can be obtained from the solution for pwmm by taking the d2-brane limit .
after the redefinitions @xmath108 , d2-brane limit is written as @xmath109 indeed , in this limit , fig.1 becomes fig.2 ( left ) and the background potential for pwmm becomes @xmath110 one can neglect the first term since it does not affect the gravity solution which depends only on @xmath111 , @xmath112 , @xmath113 and @xmath114 .
thus , the background potential for pwmm exactly reduces to that for sym on @xmath4 in the limit . by taking the d2-brane limit of the integral equation and the potential ,
we obtain @xmath115 and @xmath116 respectively . here
@xmath117 is a constant potential on the @xmath35-th disk , @xmath118 .
note that @xmath37 and @xmath117 are determined by @xmath105 and .
the solution of gives a general solution to the electrostatic problem for sym on @xmath4 .
the electrostatic system associated with the gravity dual of @xmath1 sym on @xmath2 is shown in fig.2 ( right ) . the case for the trivial vacuum
was studied in @xcite . here
, we generalize their result .
this can be obtained from that for sym on @xmath4 by compactifying the @xmath21 direction to @xmath119 with the background potential intact @xcite .
we start from the solution for sym on @xmath4 , and , with disks periodically arranged .
we change the labelling of the disks so that they are labelled by two integers @xmath120 , where @xmath121 and @xmath122 .
@xmath35 is a label of a single period and @xmath123 is that of each disk in the period .
so each period consists of @xmath124 conducting disks .
we put the position of each disk to be @xmath125 the charge @xmath126 and the radius @xmath127 of each disk is independent of @xmath35 : @xmath128 and @xmath129 .
the charge density @xmath130 on each disk should also be independent of @xmath35 : @xmath131 note that the naive substitutions of these conditions to and do not make sense because of the divergences coming from the periodicity . as remarked in @xcite , this divergence can be avoided by solving the electrostatic problem for the electric field rather than the potential .
hence , by differentiating with respect to @xmath92 and imposing the periodicity condition , one can obtain the integral equations for the charge densities @xmath132 , @xmath133 where @xmath134 the electric field generated by the conducting disks is obtained from as @xmath135
in this section , we investigate pwmm around a general vacuum as well as @xmath3 sym on @xmath4 and @xmath1 sym on @xmath2 .
first , we briefly review the results of pwmm obtained in @xcite , in which the localization was applied to a 1/4-bps sector and a matrix integral that describes this sector was obtained .
we then show that the eigenvalue densities of the matrix integral satisfy the same integral equations as the charge densities of the corresponding electrostatic system .
we also show that the same relation holds for the other gauge theories .
some properties of these theories are reviewed in appendix [ sym on s2 s3/zk ] .
we first review pwmm .
we follow the notation used in @xcite .
the action of pwmm is given in the ten - dimensional notation as , @xmath136 where @xmath137 \quad ( m\neq 1),{\nonumber \\}f_{ab}&=m \varepsilon_{abc}x_{c}-i[x_{a},x_{b } ] , \quad f_{am}=d_{a}x_m =- i[x_{a},x_m ] , \quad f_{mn}=-i[x_m , x_n ] , { \nonumber \\}d_1\psi&=\partial_1\psi - i[x_1,\psi ] , \quad d_{a}\psi=\frac{m}{8}\varepsilon_{abc}\gamma^{bc}\psi - i[x_{a},\psi ] , \quad d_m\psi =- i[x_m,\psi ] .
\label{f in pwmm}\end{aligned}\ ] ] the ranges of the indices are @xmath138 , @xmath139 and @xmath140 .
@xmath141 is the one - dimensional gauge field , @xmath142 and @xmath143 are @xmath144 and @xmath145 scalar fields and @xmath146 is a 16 component fermionic field . in the following we put the mass parameter @xmath147 .
the vacuum of pwmm is given by the fuzzy sphere solution , which is a matrix representation of @xmath13 algebra @xmath148}. \label{fuzzy sphere}\end{aligned}\ ] ] any representation of @xmath13 is an @xmath0 symmetric vacuum . the right - hand side of stands for the irreducible decomposition , where @xmath149}$ ] are generators of @xmath13 algebra in @xmath150-dimensional irreducible representation , @xmath40 are multiplicities of @xmath151}$ ] and @xmath34 is the number of different irreducible representations .
the parameters of pwmm around the vacuum are identified with those of the electrostatic system in the gravity dual as @xcite is used in different meanings in the present paper and in @xcite . ]
@xmath152 in addition , in @xcite , the case of @xmath153 was considered and @xmath26 was inferred to be related to the gauge coupling as @xmath154 where @xmath155 is a constant . in @xcite , the constant @xmath155
was determined as @xmath156 we claim that the relation ( [ v0 ] ) also holds for the theory around the general vacuum . in general , @xmath26 should be a function of @xmath157 and @xmath158 , where @xmath159 is the total matrix size @xcite .
since we are interested in the region where the supergravity approximation is valid , @xmath158 must be large .
now , let us consider the case where @xmath160 for any @xmath161 .
then , we are left with the electrostatic system made of an infinite conducting plate and a single conducting disk .
this is exactly the same situation studied in @xcite , so that @xmath26 should be given by ( [ v0 ] ) .
thus , if @xmath162 is large , the relation ( [ v0 ] ) holds .
since large @xmath163 means large @xmath158 , we conclude that @xmath164 when @xmath165 . since pwmm has one noncompact direction , in order to define the theory around a fixed vacuum precisely we have to specify the boundary condition . here
we choose the boundary condition such that all fields approach to the vacuum configuration as @xmath166 .
we also consider the t hooft limit , in which the tunneling between vacua are suppressed .
thus , the path integral with these conditions correctly defines pwmm around the fixed vacuum . in the following , we consider pwmm around a general fuzzy sphere vacuum given in . in this theory
we focus on the complex scalar @xmath167 @xmath168 is invariant under four supersymmetries ( 1/4-bps ) after the wick - rotation along @xmath169 . from a symmetry argument
, @xmath168 was found to describe the @xmath170 directions in @xcite .
so , this 1/4-bps sector is expected to correspond to the electrostatic system on the gravity side . applying the localization to operators made of @xmath168
, we obtain @xcite @xmath171 here @xmath172 is a @xmath173-independent hermitian matrix with the following block structure , @xmath174 where @xmath175 @xmath176 are @xmath177 hermitian matrices . in
the right - hand side @xmath178 stands for an expectation value with respect to the following partition function , @xmath179 where @xmath180 denotes the representation of ( [ fuzzy sphere ] ) , @xmath181 s are eigenvalues of @xmath175 and @xmath182^{\frac{1}{2}}. \label{1loopdet}\end{aligned}\ ] ] in ( [ 1loopdet ] ) , the product of @xmath183 runs from @xmath184 to @xmath185 .
@xmath186 represents that the second factor in the numerator with @xmath187 , @xmath188 and @xmath189 is not included in this product . in this subsection
, we show the results of the localization for sym on @xmath4 and sym on @xmath2 around a general vacuum , which can be obtained by applying the d2-brane limit and the t - duality to with @xcite .
sym on @xmath4 around a general monopole vacuum can be obtained from pwmm in the d2-brane limit .
the d2-brane limit in the pwmm language reads @xmath190 in this limit , with becomes , up to an overall constant , @xmath191^{\frac{1}{2 } } \nonumber\\ & \times \prod_{s , t=1}^{\lambda } \prod_{\substack{j=|q_s - q_t| \\
j\neq 0}}^{\infty } \prod_{i=1}^{n_2^{(s ) } } \prod_{j=1}^{n_2^{(t ) } } \left [ \frac{\left\{1+\left(\frac{m_{si}-m_{tj}}{2j+2}\right)^2\right\ } \left\{1+\left(\frac{m_{si}-m_{tj}}{2j}\right)^2\right\ } } { \left\{1+\left(\frac{m_{si}-m_{tj}}{2j+1}\right)^2\right\}^2 } \right]^{\frac{1}{2 } } e^{-\frac{m\pi } { g_{s^2}^2}\sum_{s , i } m_{si}^2 } , \label{matrix model s2}\end{aligned}\ ] ] where @xmath192 is the vandermonde determinant .
the parameters of sym on @xmath4 around correspond to those in the gravity dual as @xmath193 sym on @xmath2 around a general vacuum characterized by the holonomy can be obtained by applying the t - duality to sym on @xmath4 .
the t - duality can be realized in the same manner as in section [ electrostatic problem for sym on rxs3/zk ] . in , we first replace @xmath194 where @xmath195 and @xmath196 .
next , we set @xmath197 and then drop the @xmath35-dependence of @xmath198 and @xmath199 : @xmath200 and @xmath201 .
finally , up to an overall constant , we end up with @xmath202^{\frac{1}{2 } } \nonumber\\ & \times \prod_{u=-\infty}^{\infty}\prod_{\alpha,\beta } \prod_{\substack{j=|ku/2+(\alpha-\beta)/2|\\j\neq 0}}^{\infty } \prod_{i=1}^{n_2^{(\alpha)}}\prod_{j=1}^{n_2^{(\beta ) } } \left [ \frac{\left\ { 1+\left(\frac{m_{\alpha i}-m_{\beta j}}{2j+2 } \right)^2 \right\ } \left\ { 1+\left(\frac{m_{\alpha i}-m_{\beta j}}{2j}\right)^2 \right\ } } { \left\{1+\left(\frac{m_{\alpha i}-m_{\beta j}}{2j+1}\right)^2\right\}^2 } \right]^{\frac{1}{2 } } { \nonumber \\}&\times e^{-\frac{4\pi^2}{k g_{s^3/z_k}^2}\sum_{\alpha , i}m_{\alpha i}^2 } , \label{matrix model s3/zk}\end{aligned}\ ] ] where @xmath203 run over the elements of @xmath204 , @xmath205 is the vandermonde determinant .
we have also used @xmath206 now , we evaluate the matrix integral with in the regime where in the gravity side the classical gravity approximation is valid .
this regime corresponds in the gauge theory side to the t hooft limit @xmath207 and @xmath208 for arbitrary @xmath35 .
the t hooft limit suppresses the bulk string coupling while the conditions make the @xmath209 corrections negligible .
note that @xmath210 is the square of the @xmath31 radius in the @xmath35-th ns5-brane throat @xcite . also , as shown in appendix c , the second condition of means that the radius of @xmath31 near the tip of a disk in the electrostatic system is large . in these limits , and , one can evaluate by applying the saddle point approximation , which becomes exact in these limits . when @xmath211 , one can rewrite the measure factor in as @xmath212 for @xmath187 , and @xmath213 for @xmath214 , where `` @xmath215 '' stands for @xmath216 corrections .
we introduce the eigenvalue densities defined for each @xmath35 as @xmath217 in the large @xmath40 limit , @xmath218 s become continuous functions .
then , we obtain the effective action for with @xmath219 \rho^{(s)}(x)\rho^{(t)}(y ) { \nonumber \\}&-\sum_{s=1}^{\lambda}\mu_s\left(\int dx \rho^{(s)}(x)-n_2^{(s)}\right ) , \label{effective action}\end{aligned}\ ] ] where @xmath220 s are the lagrange multipliers for the normalization of @xmath218 s .
we assume that @xmath218 has its support on @xmath221 $ ] . as shown in appendix c , in the limit of and the extents of @xmath218
become large ; @xmath222 .
using the fact that @xmath223 can be approximated to @xmath224 as @xmath225 , we obtain the following saddle point equations @xmath226\rho^{(t)}(u ) { \nonumber \\}&=\frac{\mu_s}{\pi}-\frac{2d_s}{\pi g^2}x^2 , \label{saddle eq}\end{aligned}\ ] ] where @xmath227 and @xmath220 are determined from @xmath228 notice that the saddle point equations of the eigenvalue densities take a very similar form as the integral equations for the charge densities . in fact , by using the relations ( [ d d ] ) and ( [ v0 ] ) , one can find that they are exactly the same equations .
thus , we arrive at the relations @xmath229 and @xmath230 namely , the eigenvalue density on the gauge theory side has exactly the same functional form and parameter dependence as the charge density on the gravity side , up to the trivial rescaling . hence , they can naturally be identified with each other and this identification relates the degrees of freedom on the gauge theory side to the background geometry on the gravity side . by integrating both sides of over @xmath221 $ ] and using and , we find that those relations are consistent with and .
if one finds exact solutions of and , one can check the relations , and , more explicitly . although we could not find general exact solutions , still we can solve those equations in particular parameter regions .
if a conducting disk is isolated at a distance from the other disks , the term with an integration in the integral equation of the disk becomes negligible .
then , the solution is simply given by a quadratic function .
this is effectively the same situation as the d2-brane limit with @xmath153 considered in @xcite . in the same way
, we can consider two isolated disks which effectively form the same system as the ns5-brane limit with @xmath153 @xcite .
in these cases , one can find the exact solutions and check the relations , and , more directly .
the equivalence between the charge density and the eigenvalue density also holds for the other gauge theories with @xmath0 symmetry .
as shown in section [ localization s2 s3/zk ] , in the d2-brane limit , the partition function reduces to the matrix integral for sym on @xmath4 given by .
in addition , by taking the t - duality , we end up with the matrix integral for sym on @xmath2 given by .
if we apply the corresponding limits to the integral equation of the charge density , we obtain the integral equations and for sym on @xmath4 and sym on @xmath2 , respectively . in these cases , the integral equations and for the charge densities
can also be identified with the saddle point equations for and , respectively , where the same relations as and hold .
in this paper , we showed that the bubbling geometries in type iia supergravity are realized in the gauge theories with @xmath0 symmetry .
we found that the charge densities of the electrostatic systems in the gravity dual are equivalent to the eigenvalue densities of the matrix integrals which govern the 1/4-bps sector of the gauge theories . on the gravity side ,
the bubbling geometries are given in terms of the electrostatic potential of electrostatic systems with conducting disks .
first , we have considered the electrostatic system corresponding to pwmm around a general vacuum .
we have shown that the boundary conditions of the potential are given by a system of dual integral equations . extending the method to analyze the dual integral equations written in @xcite , we have reduced the dual integral equations to the fredholm integral equations of the second kind for the charge densities on the disks . by taking the d2-brane limit or performing the t - duality as well as the d2-brane limit , we have also obtained the same type of integral equations for the charge densities in the electrostatic system corresponding to sym on @xmath4 or sym on @xmath2 . on the gauge theory side
, we have investigated the matrix integrals that describe 1/4-bps sectors of the gauge theories .
first , we have considered the case for pwmm around a general vacuum in the regime where the supergravity approximation is valid . in this regime , we have derived the saddle point equations of the eigenvalue densities of the matrix integral , which are almost the same integral equations for the charge densities on the gravity side .
then we have found that under the identifications of and the integral equations of the eigenvalue densities are exactly equivalent to those for the charge densities . as the d2-brane limit and the t - duality of pwmm lead to the other gauge theories with @xmath0 symmetry , that is , sym on @xmath4 and sym on @xmath2 ,
we have also shown the equivalence of the charge densities and the eigenvalue densities .
thus , we have concluded that since the bubbling geometries are completely determined by the charge densities the geometries are constructed from the eigenvalue densities of the gauge theories with @xmath0 symmetry .
finally , let us comment on iia lst on @xmath14 , which is another theory with @xmath0 symmetry .
like other @xmath0 symmetric theories , lst on @xmath14 is thought to have many discrete vacua and for each vacuum there exists a gravity dual given by type iia bubbling geometry @xcite .
the gravity dual of lst around the trivial vacuum was elaborated in @xcite and shown to be obtained from a double scaling limit of the gravity dual of pwmm around a particular vacuum .
although it is also expected that the gravity dual of lst around a general vacuum can be obtained from the same kind of double scaling limit of the gravity dual of pwmm , some careful analysis seems to be needed .
so , we will return to this issue in a separate paper .
the work of y.a . and s.s .
is supported in part by the jsps research fellowship for young scientists .
the work of s.s is also supported in part by the jsps grant - in - aid for scientific research ( a ) no .
let @xmath231 be a function defined on @xmath232 .
we define two functions on @xmath232 using @xmath231 as @xmath233 where @xmath123 is a positive half integer , @xmath234 is a given weight function on @xmath232 and @xmath235 is the bessel function of the first kind of order @xmath236 defined by @xmath237 in this appendix , we consider a problem of finding a solution for @xmath231 , which solves the following equations called the dual integral equations .
@xmath238 here we have divided the positive real line @xmath232 to two segments denoted by @xmath239 and @xmath240 , which are written as @xmath241 and @xmath242 , respectively . @xmath243 and @xmath244 are assumed to be known functions defined on @xmath239 and @xmath240 , respectively .
we will see that the problem reduces to a problem of solving a single fredholm integral equation of the second kind @xcite . in the following , for any function @xmath245 on @xmath232 ,
we denote by @xmath246 and @xmath247 the restrictions of @xmath245 to @xmath239 and @xmath240 , respectively .
we assume that @xmath248 unless @xmath249 ( @xmath250 ) , so that the original function can be written as @xmath251 .
for example , the equation ( [ phi f ] ) can be written in this notation as @xmath252 for @xmath253 .
we also introduce the modified hankel transformation , @xmath254 the inverse transformation is given by @xmath255 it is easy to see that @xmath256 where @xmath257 is defined by @xmath258 and the dot in ( [ phi 2 ] ) denotes the product of functions defined as usual by @xmath259 .
we first put @xmath260 and substitute this to ( [ chi 2 ] ) . then , after performing the inverse transformation ( [ inverse hankel ] ) twice , we obtain @xmath261 where we have defined @xmath262 one can show that the transformation @xmath263 can be written as @xmath264 see @xcite for the definition for @xmath265 . from ( [ h and chi ] ) and ( [ def of k ] )
, we find that the solution for @xmath266 is given by @xmath267 similarly , by applying the same calculation to ( [ phi 2 ] ) , one can obtain the following equation for @xmath268 .
@xmath269 the function @xmath270 is defined by ) is obtained by using ( [ h2 and g ] ) and the relation , @xmath271 . ] @xmath272 where @xmath273 is defined by @xmath274 and it is written more explicitly as @xmath275 note that @xmath270 depends only on the known functions @xmath276 , @xmath243 and @xmath244 . for the second term in the left - hand side of ( [ eq for h1 ] )
, we interchange the order of the integration , so that @xmath277 where the integral kernel @xmath278 is defined by @xmath279 hence , we conclude that @xmath280 is the solution of the fredholm integral equation of the second kind , @xmath281 the two equations ( [ h2 and g ] ) and ( [ fredholm ] ) fully determines @xmath282 . now , we consider a generalization of this result to vector - valued functions .
this is easy since the above problem is linear in @xmath231 .
let us consider a set of functions @xmath283 which are determined by the equations , @xmath284 where @xmath285 and @xmath286 are now defined by @xmath287 for each @xmath35 , @xmath288 and @xmath289 are assumed to be known functions and @xmath290 are the two connected intervals , the sum of which is equal to @xmath232 . if we write for each @xmath35 @xmath291 it is easy to see that @xmath81 are determined by the following equations .
@xmath292 where @xmath293 and @xmath294 are defined by @xmath295 and @xmath296
in this appendix , we review sym on @xmath4 and sym on @xmath2 . see for detail @xcite .
the action of sym on @xmath4 is given by @xmath297)^2-\frac{1}{2}(d_1x_m)^2 { \nonumber \\}&\hspace{35 mm } -\frac{1}{4}(2\varepsilon_{abc}x_{c}+il^{(0)}_{a}x_{b}-il^{(0)}_{b}x_{a } -i[x_{a},x_{b}])^2 { \nonumber \\}&\hspace{35 mm } + \frac{1}{2}(l_a^{(0)}x_m-[x_a , x_m])^2-\frac{1}{2}x_m^2+\frac{1}{4}[x_m , x_n]^2 { \nonumber \\}&\hspace{35 mm } -\frac{i}{2}\psi \gamma^1 \partial_1 \psi + \frac{1}{2}\psi\gamma^{a } l^{(0)}_{a } \psi-\frac{3i}{8}\psi\gamma^{234}\psi -\frac{1}{2}\psi\gamma^m[x_m,\psi ] \bigr ) , \label{sym on rxs2 1}\end{aligned}\ ] ] where @xmath298 are ordinary angular momentum operators and the radius of @xmath30 is set to be @xmath299 @xcite .
@xmath141 is the gauge field corresponding to the @xmath300 direction , @xmath143 are the scalar fields and @xmath146 is a ten - dimensional majorana - weyl spinor with 16 components . @xmath142 include gauge fields and a scalar field : @xmath301 where @xmath302 , @xmath303 , @xmath304 and @xmath305 .
@xmath306 and @xmath307 are the gauge fields in the local lorentz frame and @xmath308 is the scalar field .
the vacuum of this theory is given by the dirac monopole @xmath309 where @xmath310 are monopole charges .
the upper and lower signs represent the northern region where @xmath311 and the southern region where @xmath312 on @xmath30 , respectively .
the localization can be applied to the following scalar field sitting at @xmath313 , @xmath314 this scalar field preserves the same supersymmetries as .
note that @xmath315 .
the vev of operators made of can be computed in terms of the matrix integral obtained in the d2-brane limit of with .
@xmath316 is the @xmath317-orbifold of the round @xmath32 along the @xmath119-fiber direction .
when @xmath32 is parametrized by the coordinates @xmath318 ( @xmath319 ) with the metric @xmath320 @xmath316 is realized as the following coordinate identification ; @xmath321 . then , sym on @xmath2 is obtained by keeping the modes of fields of sym on @xmath322 that are invariant under the @xmath317 action .
the action of sym on @xmath2 takes the form , @xmath323)^2-\frac{1}{2}(d_1x_m)^2 { \nonumber \\}&\hspace{35 mm } -\frac{1}{4}(2\varepsilon_{abc}x_{c}+i{\mathcal{l}}_{a}x_{b}-i{\mathcal{l}}_{b}x_{a}-i[x_{a},x_{b}])^2 { \nonumber \\}&\hspace{35 mm } + \frac{1}{2}({\mathcal{l}}_ax_m-[x_a , x_m])^2-\frac{1}{2}x_m^2+\frac{1}{4}[x_m , x_n][x^m , x^n ] { \nonumber \\}&\hspace{35 mm } -\frac{i}{2}\psi \gamma^1 \partial_1 \psi + \frac{1}{2}\psi\gamma^{a } { \mathcal{l}}_{a } \psi-\frac{3i}{8}\psi\gamma^{234}\psi-\frac{1}{2}\psi\gamma^m[x_m,\psi ] \bigr ] , \label{sym on rxs3/zk}\end{aligned}\ ] ] where @xmath324 are the killing vectors on @xmath316 @xcite . @xmath141 and @xmath142 are the gauge fields for @xmath300 and @xmath316 directions , respectively , @xmath143 are the scalar fields and @xmath146 is a ten - dimensional majorana - weyl spinor with 16 components .
the vacuum of this theory is parametrized by the holonomy around the @xmath119-fiber direction of @xmath316 .
the holonomy can be labeled by @xmath325 where @xmath326 and @xmath327 are the multiplicities of the holonomy phase @xmath328 .
the total matrix size is given by @xmath329 .
note that when @xmath330 this theory is the sym on @xmath322 and has the unique and trivial vacuum .
the localization can be applied to the following wilson loop operator defined at a fixed @xmath173 , @xmath331 where @xmath332 .
this also preserves the same supersymmetries as .
the correlation functions of can be computed in terms of the matrix integral obtained by applying the t - duality to .
in this appendix , we show that @xmath333 is a sufficient condition for the large @xmath31 radius at the tips of the disks in the electrostatic problem . at the tip of a disk , the disk radius @xmath300 and the radius @xmath334 of @xmath31 are related as @xcite @xmath335 in the string unit , @xmath336 .
then , under the identification ( [ x r ] ) , the @xmath31 radius is large if and only if @xmath222 . in the following ,
we show that @xmath222 if @xmath333 .
we assume that the index @xmath35 labels the disks in the order of the @xmath21-coordinate , namely , @xmath337 ( @xmath338 ) .
first we divide the theory described by ( [ effective action ] ) into three parts .
the first is the free part , the action of which is given by @xmath339 the second is the self - interaction part given by @xmath340 the third is the interaction between different @xmath35 and @xmath341 , defined by @xmath342 \rho^{(s)}(x)\rho^{(t)}(y ) .
\label{s3}\end{aligned}\ ] ] the total theory is described by the sum of these .
but for the moment , let us consider more generally the theory defined by @xmath343 , where @xmath123 and @xmath344 are parameters .
we start with the simplest free theory with @xmath345 . in this case
, the extents of the eigenvalues can be easily estimated as @xmath346 .
this gives a typical length scale of the free theory . from ( [ s2 ] ) and ( [ s3 ] ) , one can also read off the typical length scale of the interaction potentials . for the self - interaction ,
it is given by @xmath347 , where @xmath348 denotes the separation distance between two eigenvalues .
for the interaction between different @xmath35 and @xmath341 , the scale ( for a fixed @xmath35 ) is equal or greater than @xmath349 , namely , @xmath350 .
the lower bound is saturated by the interaction between @xmath35 and @xmath351 .
then , let us consider turning on the interactions to recover the theory with @xmath352 .
the typical scale @xmath353 of the free theory should be modified by the interactions , which have structures with the length scale equal or greater than @xmath349 ( note that we always assume that @xmath354 ) .
the modified scale should be at least greater than @xmath355 , since there is nothing which provides a finer scale than these . if the modified scale is @xmath356 , this is always large enough when @xmath354 .
if the modified scale is @xmath357 , this is large if @xmath358 .
therefore , we conclude that if @xmath358 , the typical extents of the eigenvalues are always much greater than 1 .
i. r. klebanov , in * trieste 1991 , proceedings , string theory and quantum gravity 91 * 30 - 101 and princeton univ
. - pupt-1271 ( 91/07,rec.oct . )
72 p [ hep - th/9108019 ] .
p. h. ginsparg and g. w. moore , in * boulder 1992 , proceedings , recent directions in particle theory * 277 - 469 . and yale univ .
new haven - yctp - p23 - 92 ( 92,rec.apr.93 ) 197 p. and los alamos nat . lab .
- la - ur-92 - 3479 ( 92,rec.apr.93 ) 197 p [ hep - th/9304011 ] .
s. mukhi , hep - th/0310287 .
h. lin , o. lunin and j. m. maldacena , jhep * 0410 * , 025 ( 2004 ) [ hep - th/0409174 ] . s. yamaguchi , int .
j. mod .
a * 22 * , 1353 ( 2007 ) [ hep - th/0601089 ] .
d. berenstein , jhep * 0407 * , 018 ( 2004 ) [ hep - th/0403110 ] . y. takayama and a. tsuchiya , jhep * 0510 * , 004 ( 2005 ) [ hep - th/0507070 ] .
h. mori and s. yamaguchi , arxiv:1404.0930 [ hep - th ] .
y. asano , g. ishiki , t. okada and s. shimasaki , jhep * 1405 * , 075 ( 2014 ) [ arxiv:1401.5079 [ hep - th ] ] .
h. lin and j. m. maldacena , phys .
d * 74 * , 084014 ( 2006 ) [ hep - th/0509235 ] .
h. ling , a. r. mohazab , h. -h .
shieh , g. van anders and m. van raamsdonk , jhep * 0610 * , 018 ( 2006 ) [ hep - th/0606014 ] .
v. pestun , commun .
phys . * 313 * , 71 ( 2012 ) [ arxiv:0712.2824 [ hep - th ] ] .
y. asano , g. ishiki , t. okada and s. shimasaki , jhep * 1302 * , 148 ( 2013 ) [ arxiv:1211.0364 [ hep - th ] ] .
j. maldacena , m. m. sheikh - jabbari and m. van raamsdonk , jhep * 0301 * ( 2003 ) 038 [ arxiv : hep - th/0211139 ] .
m. berkooz , m. rozali and n. seiberg , phys .
b * 408 * , 105 ( 1997 ) [ hep - th/9704089 ] .
n. seiberg , phys .
b * 408 * , 98 ( 1997 ) [ hep - th/9705221 ] .
o. aharony , class .
quant . grav .
* 17 * , 929 ( 2000 ) [ hep - th/9911147 ] .
d. kutasov , `` introduction to little string theory , '' superstrings and related matters .
proceedings , spring school , trieste , italy , april 2 - 10 , 2001 g. ishiki , s. shimasaki , y. takayama and a. tsuchiya , jhep * 0611 * ( 2006 ) 089 [ arxiv : hep - th/0610038 ] .
w. taylor , phys .
b * 394 * , 283 ( 1997 ) [ hep - th/9611042 ] .
h. ling , h. -h .
shieh and g. van anders , jhep * 0702 * , 031 ( 2007 ) [ hep - th/0611019 ] . | we study the gauge / gravity duality between bubbling geometries in type iia supergravity and gauge theories with @xmath0 symmetry , which consist of @xmath1 super yang - mills on @xmath2 , @xmath3 super yang - mills on @xmath4 and the plane wave matrix model .
we show that the geometries are realized as field configurations in the strong coupling region of the gauge theories . on the gravity side ,
the bubbling geometries can be mapped to electrostatic systems with conducting disks .
we derive integral equations which determine the charge densities on the disks . on the gauge theory side ,
we obtain a matrix integral by applying the localization to a 1/4-bps sector of the gauge theories .
the eigenvalue densities of the matrix integral turn out to satisfy the same integral equations as the charge densities on the gravity side .
thus we find that these two objects are equivalent .
[ cols= " < " , ] + emergent bubbling geometries in + gauge theories with @xmath0 symmetry yuhma asano@xmath5 , goro ishiki@xmath6 , and shinji shimasaki@xmath7 + @xmath5 _ department of physics , kyoto university _ + _ kyoto , 606 - 8502 , japan _
+ @xmath8 _ yukawa institute for theoretical physics , kyoto university _ + _ kyoto , 606 - 8502 , japan _
+ @xmath9 _ theory center , high energy accelerator research organization ( kek ) _
+ _ ibaraki 305 - 0801 , japan _ + |
recently , there have been considerable developments as regards large field inflation with plateau - like inflaton potentials , driven by the models compatibility with observational data @xcite and their ties to supergravity and string theory @xcite@xcite . on the theoretical side it has , for example ,
been shown that such models naturally arise from phenomenological supergravity .
the underlying hyperbolic geometry of the moduli space and the flatness of the khler potential in the inflation direction is conveniently described in terms of a field variable @xmath4 , which gives rise to a kinetic term in the lagrangian with a pole at the boundary of the moduli space . by making a transformation to a canonical variable @xmath5 ,
the moduli space near its boundary becomes stretched , leading to an inflationary potential @xmath6 with an asymptotic plateau - like form for @xmath6 .
this stretching results in predictions that are quite insensitive to the original form of the potential @xmath7 . in particular this leads to universal properties in @xmath8 diagrams , which motivates calling these models @xmath0-attractors , where @xmath0 is a constant parameter describing the exponential asymptotic flattening of the potential @xmath6 in the einstein frame .
it should be noted that this theoretical perspective unifies and contextualizes results for several previous models such as the starobinski and the higgs inflation models . for additional intriguing aspects such as couplings to other fields with couplings which become exponentially small for large fields @xmath5 , as well as further background discussions ,
see @xcite@xcite .
we take the einstein frame formulation as our starting point and restrict the discussion to the flat , spatially homogeneous , isotropic friedmann - lematre - robertson - walker ( flrw ) spacetimes .
we thereby consider the following field equations for a canonically normalized inflaton field @xmath5 with a potential @xmath6 : . ]
[ fieldeq ] @xmath9 where @xmath10 ; an overdot signifies the derivative with respect to synchronous proper time , @xmath11 ; @xmath12 is the hubble variable , where @xmath13 is the cosmological scale factor , with evolution equation @xmath14 , which decouples from the above equations .
furthermore , our focus will be on @xmath1- and @xmath2-models defined by the potentials [ pot ] @xmath15 with @xmath16 , respectively ( see fig .
[ fig : pot ] ) .
the modest aim in this paper is to present a dynamical systems formulation that allows us to give a complete global classical description of the solution spaces of the above models , including asymptotic properties , illustrated with pictures describing compactified two - dimensional state spaces .
our motivation for this is threefold : * as will be discussed , the above equations and potentials allow one to rather directly get a feeling for the solution space and its properties , but all aspects are not obvious .
it is therefore of value to show how a complete understanding can be obtained in a rigorous manner , especially since this exemplifies how one can address other reminiscent problems .
* by reformulating the problem as a regularized dynamical system on a reduced compactified state space , we provide an example that helps give the dynamical systems community access to inflationary cosmology .
this also makes powerful mathematical tools available to an area where such methods have not , in our opinion , been used to full effect .
for example , as will be shown , the universal inflationary properties of the models are intimately connected with certain aspects concerning center manifolds , which also result in approximation methods complementing heuristic methods such as the slow - roll approximation . *
the present dynamical systems formulation can be adapted so that it can be used in more general contexts such as anisotropic spatially homogeneous cosmology and even for models without any symmetries at all , as shown for perfect fluids in e.g. , refs .
@xcite and @xcite .
such an extension of the present paper would make it possible to address the issue of initial conditions for inflation in a generic context .
let us now turn to the system . by treating @xmath17 as an independent variable ,
we obtain a reduced state space ( because the equation for the scale factor @xmath13 decouples ) described by the state vector @xmath18 obeying the constraint ; i.e. , the problem can be regarded as a two - dimensional dynamical system .
the @xmath1- and @xmath2-models are examples of models with a non - negative potential @xmath6 with a single extremum point , a minimum , conveniently located at @xmath19 , for which @xmath20 . as a consequence
these models admit a minkowski ( fixed point / critical point / equilibrium point ) solution at @xmath21 .
due to , all other solutions either have a positive or negative @xmath22 .
since we are interested in cosmology we consider @xmath23 .
it follows that @xmath22 is monotonically decreasing since eq .
yields @xmath24 , where @xmath25 requires @xmath26 ; moreover , @xmath17 can not remain zero since @xmath26 implies that @xmath27 , where @xmath28 since we exclude the only extremum point at @xmath29 by assuming that @xmath23 .
thus , the graph of @xmath22 just passes through an inflection point when @xmath26 .
the system can be discussed in terms of the following heuristic picture : equation can be interpreted as an equation for a particle of unit mass with a one - dimensional coordinate @xmath5 , moving in a potential @xmath6 with a friction force @xmath30 , while @xmath31 can be viewed as a monotonically decreasing energy in the constraint .
running backwards in time leads to a particle with ever - increasing energy moving in the potential @xmath6 , where @xmath32 . in combination with the shape of a given potential , this yields an intuitive picture of the dynamics
let us now turn to the specific potentials .
in the case of the @xmath1- and @xmath2-models both potentials behave as @xmath33 for small @xmath5 .
the potential for the @xmath1-models ( @xmath2-models ) has a plateau given by @xmath34 when @xmath35 ( @xmath36 ) , while @xmath37 behaves as @xmath38 when @xmath39 .
in contrast to the @xmath1-models , the potential , and thereby the field equations , of the @xmath2-models exhibits a discrete symmetry under the transformation @xmath40 .
let us begin our heuristic discussion of the dynamics of the @xmath1- and @xmath2-models with dynamics toward the future .
in both cases , if a solution has obtained an `` energy '' @xmath41 it follows straightforwardly that the solution will end up at the minkowski fixed point .
but does it do so by just `` gliding '' down the potential , or does it do so by damped oscillations , i.e. , is the motion of @xmath5 overdamped or underdamped ?
this does not follow from the heuristic particle discussion , but requires further analysis .
attempting to do so by solving the constraint for @xmath22 , i.e. , by setting @xmath42 in , leads to an unconstrained two - dimensional dynamical system for @xmath43 .
this system , however , has an unbounded state space and differentiability problems at @xmath44 for potentials that behave as @xmath45 for small @xmath5 .
these problems are of course not insurmountable , but they prevent global pictures of the solution spaces that accurately reflect the asymptotic features of the solutions . nevertheless , it is not difficult to show that the motion is underdamped and that solutions with @xmath41 undergo damped oscillations .
but do all solutions end at the minkowski state , or are there solutions that asymptotically end with an energy @xmath34 at infinitely large @xmath5 ?
in other words , is the minkowski state a global future attractor ? ( as we will see , the answer is `` yes '' for the present potentials . ) one would also perhaps guess that there is a single solution that begins at an infinite value of the scalar field with an initial energy @xmath46 , corresponding to an asymptotic de sitter state , with the solution initially slowly rolling down the potential , which , as we will show , is indeed the case .
what about the past dynamics ? from an inflationary point of view one would perhaps argue that solutions are only physically acceptable with initial data for which @xmath47 .
however , all solutions , except for the single one coming from the de sitter plateau with an initially infinite scalar field , originate from @xmath32 . one
might then take the viewpoint that the previous evolution of solutions before @xmath47 is physically irrelevant .
however , all this assumes that the inflationary scenario is correct and prevents investigations that either justify this or cast doubt on it .
moreover , even if this is assumed to be correct , one still wants approximations for the solutions before the quasi - de sitter stage , as done in , e.g. , ref .
@xcite , where a massless state is used for this purpose .
but this is intimately connected with the limit @xmath32 , so let us therefore consider this limit from the present heuristic perspective . because the potential is symmetric and bounded , the heuristic description of the dynamics of the @xmath2-models is simpler than that for the @xmath1-models .
we therefore begin with the @xmath2-models .
since the potential in this case is bounded , @xmath48 , it follows from that the limit @xmath32 implies that @xmath49 and hence that the influence of the potential on the dynamics becomes asymptotically negligible ; i.e. , the asymptotic state is expected to be that of a massless scalar field .
furthermore , we expect two physically equivalent representations associated with @xmath50 toward the past . for the @xmath1-models
the potential is no longer bounded and this results in a somewhat more complicated situation . nevertheless , toward the past , irrespective of the value of @xmath0
, we would expect an open set of models to approach a massless state toward @xmath35 for similar reasons as for the @xmath1-models .
furthermore , we expect that whether or not all solutions end up there depends on the steepness of the potential in the limit @xmath39 . in this limit
the potential behaves as an exponential , and problems with an exponential can be viewed as scattering a particle against a potential wall if the potential is steep enough ; in such a situation we expect that the massless state with @xmath35 describes the past asymptotic behavior of all solutions . if the potential is not steep enough to accomplish this the situation becomes more complicated .
nevertheless , based on the knowledge of the dynamics of a single exponential one might guess that there is an open set of solutions that approaches a massless state at @xmath51 and that there exists a single solution that approaches this limit in a power - law fashion .
moreover , if the steepness of the exponential limit is sufficiently moderate , we expect this solution to describe an early inflationary power - law state .
although quite helpful , the above heuristic considerations lead to a rather scattered impression of the global dynamics and leave some remaining unanswered questions .
in addition , we have not obtained any quantitative results , and , what is even worse , the heuristic picture runs into increasing problems when trying to use it in more general contexts involving additional degrees of freedom .
in contrast , the dynamical systems formalism presented below is applicable to more general situations than the present one , which rather acts as a pedagogical example . with this in mind
we will now develop a formalism that addresses the above issues and at a glance describes the global situation rigorously , and also makes techniques available for describing quantitatively correct approximations for solutions in the past and future regimes ( indeed , they can be combined to even give global approximations for the solutions to arbitrary accuracy if one is so inclined , see @xcite ) . to avoid differentiability problems and to obtain asymptotic approximations and a global picture of the solution space , we change variables .
we do so by following the treatment of monomial potentials @xmath52 in @xcite and @xcite , but adapting the formulation to the particular features of the potentials and .
we derive three complementary dynamical systems since the global one is not optimal for quantitative descriptions in all parts of the state space ; instead the global picture should be seen as a collecting ground which gives the overall picture .
since the dependent variables are defined in a similar manner for @xmath1- and @xmath2-models , while the independent variables differ , we define the three complementary sets of dependent variables in this section , while the independent variables and the dynamical systems and their analysis will be presented in the subsequent two sections .
we begin by defining the hubble - normalized or , equivalently in the present flat flrw cases , energy density - normalized dimensionless variables ( thereby capturing the physical essence of the problem ) , by making the following variable transformation , @xmath53 is used because @xmath54 , in a multidimensional kaluza - klein perspective , is analogous to hubble - normalized shear in anisotropic cosmology , where @xmath55 is standard notation ; see e.g. ref .
@xcite . ] : [ vardef ] @xmath56 where @xmath57 takes the following explicit form for the @xmath1- and @xmath2-models , @xmath58 respectively .
the quantity @xmath59 in represents the number of @xmath3-folds with respect to some reference time @xmath60 where @xmath61 .
below , for simplicity , we assume that @xmath62 is a positive integer . since @xmath22 is monotonically decreasing it follows that @xmath63 is monotonically increasing .
furthermore , when @xmath64 , @xmath65 , and @xmath66 . with the above definitions , the gauss constraint takes the form @xmath67 the state space thereby has a cylinderlike structure ( cylinder structure when @xmath68 ) . the constraint can be solved globally by introducing an angular variable @xmath69 according to [ thetadef ] @xmath70 which leads to an unconstrained dynamical system for the state vector @xmath71 . for future purposes we note that @xmath72 ( with @xmath73 when @xmath68 ) , and @xmath74 to obtain a bounded ( relatively compact ) state space with state vector @xmath75 , we make the transformation @xmath76 thus
, @xmath77 is monotonically increasing where @xmath78 , @xmath79 , and @xmath80 .
the deceleration parameter , @xmath81 , is defined and given by @xmath82 to proceed with the choice of independent variables , we treat the @xmath1- and @xmath2-models separately in the following two sections , where we also perform a complete local and global dynamical systems analysis of these models .
we end the paper with some concluding remarks in sec .
[ sec : concl ] , e.g. about the relationship between the center manifold analysis performed in the two @xmath1- and @xmath2-model sections and the slow - roll approximation .
using the dependent variables given in eq . for the @xmath1-models and @xmath83 as the independent variable , where @xmath84 results in the following evolution equations for the state vector @xmath85 , [ dynsys1e ] @xmath86 and the constraint @xmath87 where it has been convenient to define @xmath88 the state space is bounded by the conditions that @xmath89 and that @xmath90 since @xmath91 it follows that the physical state space is bounded toward the past by the invariant subsets @xmath92 for @xmath93 and @xmath94 for @xmath95 ( recall that @xmath63 is monotonically increasing toward the future and therefore decreasing toward the past ) .
their intersection , @xmath96 , corresponds to two fixed points @xmath97 .
furthermore , the @xmath94 subset is divided into two disconnected parts by a fixed point @xmath98 located at @xmath99 . due to the regularity of the above equations we can include the invariant boundary , ( @xmath92 for @xmath93 ) @xmath100 ( @xmath94 for @xmath95 ) , which we refer to as the past boundary . indeed
, it is necessary to include this boundary in order to describe past asymptotics since , as will be shown , all solutions originate from the fixed points on this boundary .
note further that the equations on the @xmath92 subset are identical to those for an exponential potential @xmath101 since in this case @xmath102 moreover , the equations on the @xmath103 subset are identical to those that correspond to a constant potential , which can be seen by setting @xmath104 in the above equation . by globally solving the constraint by using eq .
we obtain the following unconstrained dynamical system [ dynsys2e ] @xmath105 finally , by using @xmath77 and @xmath69 and changing the time variable from @xmath106 to @xmath107 according to @xmath108 we obtain the regular dynamical system [ dynsys3e ] @xmath109g.\end{aligned}\ ] ] apart from including the past boundary , which in the present variables is given by @xmath110 when @xmath111 and @xmath112 for @xmath113 , we also include the future boundary @xmath114 , which corresponds to @xmath115 and the final minkowski state .
thus , the resulting extended state space is given by a finite cylinder with the region @xmath116 with @xmath117 removed ( see fig .
[ fig : estatespace ] ) . from the definitions , and the fact that @xmath22 is monotonically decreasing
, it follows that @xmath63 and @xmath77 are monotonically increasing .
this is also seen in eqs . , , and ,
although further insight is gained by considering how @xmath63 , and hence @xmath77 , behaves when @xmath118 , where @xmath119 : @xmath120 since @xmath121 when @xmath122 is even and @xmath123 when @xmath122 is odd in the physical state space , it follows that by viewing the above as the coefficients in a taylor expansion , @xmath63 , and hence @xmath77 , is monotonically increasing , although the graphs of @xmath63 and @xmath77 go through inflection points when @xmath119 .
furthermore , since @xmath124 it follows that @xmath69 is monotonically decreasing at @xmath119 and thus that the solution curves in the @xmath125 state space become horizontal in @xmath77 at @xmath119 ( see fig . [
fig : esolspace ] ) .
the monotonicity properties of @xmath77 show that there are no fixed points or recurring orbits in the physical state space ( i.e. , the extended state space with the future and past invariant boundaries excluded ) .
all orbits originate from the past boundary and end at the future boundary at @xmath114 , where @xmath126 i.e. , @xmath114 corresponds to a periodic orbit [ i.e. , a periodic solution trajectory to the dynamical system ] , @xmath127 , with monotonically decreasing @xmath69 , and , hence , to a limit cycle for all solutions in the physical state space .
the structure on the past boundary is easily found since it consists of two parts , one corresponding to an exponential potential and one to a constant potential .
it consists of fixed points and heteroclinic orbits ( orbits that originate and end at distinct fixed points ) that join them .
since there are no heteroclinic cycles on the past boundary , it follows that all interior physical orbits originate from the fixed points , which are given by [ fixedpoints ] @xmath128 where @xmath129 only exists on the extended physical state space if @xmath130 .
this fixed point corresponds to the self - similar solution for an exponential potential , which yields a power - law solution , explaining the nomenclature . if @xmath131 this solution is accelerating since @xmath132 for @xmath129 .
the nomenclature @xmath97 corresponds to a massless scalar field with @xmath133 ( i.e. , it corresponds to setting the potential to zero ) , where @xmath97 implies @xmath134 , due to .
finally , @xmath98 stands for de sitter since @xmath119 for this fixed point , although note that this de sitter state corresponds to @xmath135 ; i.e. , it is an asymptotic state and not a physical de sitter solution with finite constant @xmath5 .
a local analysis of the fixed points shows that if @xmath136 then both @xmath137 and @xmath138 are sources while @xmath129 is a saddle with a single solution entering the physical state space . if @xmath139 then @xmath138 is a source , while @xmath137 is a saddle from which no solutions enter the physical state space .
arguably , the most interesting fixed point is @xmath98 , which is a center saddle , with a one - dimensional center manifold corresponding to the `` attractor solution '' or the `` inflationary trajectory '' . to describe this interior state space solution , which originates from @xmath98
, we follow @xcite and perform a center manifold analysis .
since it is more convenient to use @xmath63 than @xmath77 , we use the system , which results in the following center manifold expansion ( without loss of generality we choose @xmath140 for @xmath98 ) : @xmath141 the periodic orbit @xmath127 corresponds to a blowup of the completely degenerate minkowski fixed point in the @xmath43 formulation . as discussed in @xcite , to obtain explicit expressions for future asymptotics one can use available approximations for late stage behavior when @xmath142 , or one can use the averaging techniques developed in @xcite . the overall global solution structure for the @xmath1-models
is depicted in fig .
[ fig : esolspace ] .
note that @xmath127 is the true future attractor , and that it represents the future asymptotic behavior of _ all _ physical solutions .
finally , let us translate the above results to the original scalar field picture .
let us first consider the inflationary solution coming from the @xmath98 fixed point . in the vicinity of @xmath98 ,
@xmath143 is negative , which means that toward the past , @xmath5 is increasing .
it is not difficult to use the approximation to show that @xmath5 to leading order can be written in the form @xmath144 , where @xmath145 and @xmath146 are positive constants and where @xmath147 toward the past ; i.e. , the solution originates from @xmath148 with a subsequently slowly decreasing @xmath5 , as expected .
furthermore , note that initial data with an energy that is close to that of an initial quasi - de sitter state correspond to @xmath149 , where fig .
[ fig : esolspace ] conveniently , at a glance , shows how solutions , and their properties , are distributed in terms of such initial data . from this figure
it is obvious that there exists an open set with initial data with such initial energies that do not have a quasi - de sitter phase in their evolution .
moreover , solutions that do have an intermediate de sitter stage have solution trajectories that shadow the heteroclinic orbits @xmath150 , which correspond to scalar field models with a constant potential @xmath34 . from @xmath143
it follows directly that @xmath138 corresponds to a massless initial asymptotic state for which @xmath35 . _ if _
@xmath151 , i.e. , if @xmath152 , then @xmath137 corresponds to a massless initial asymptotic state with @xmath39 from which an open set of solutions originates .
similarly , it follows that in this case the single solution that originates from @xmath129 corresponds to an initial power - law state for which @xmath39 toward the past ; furthermore , if @xmath131 this is an initial power - law inflation state .
since @xmath129 is a saddle , it follows that there is an open set of solutions that undergo an intermediate stage of power - law inflation in this case , but note that ( i ) this stage is much shorter than the de sitter stage since @xmath129 is a hyperbolic saddle in contrast to the center saddle @xmath98 and ( ii ) this inflationary stage takes place at a much larger energy scale than that of the present quasi - de sitter inflation . on the other hand , if @xmath139 , i.e. , if @xmath153 , then all solutions originate from @xmath35 .
finally , all solutions end at the minkowski state associated with @xmath127 .
using the dependent variables given in eq . and a new time variable @xmath154 , defined by @xmath155 results in the following evolution equations for the state vector @xmath85 , [ dynscalart1 ] @xmath156 and the constraint @xmath157 where again @xmath88 the constrained dynamical system admits a discrete symmetry @xmath158 , because the potential is invariant under the transformation @xmath159 .
the state space is bounded by the conditions that @xmath89 and that @xmath160 since @xmath161 it follows that the physical state space is bounded toward the past by the invariant subset @xmath162 for @xmath163 . due to the discrete symmetry ,
this invariant subset consists of two equivalent disconnected parts , one with @xmath164 and one with @xmath165 , separated by the equivalent ( massless state ) fixed points @xmath97 at @xmath166 , @xmath167 .
the equations on the two branches of the boundary subset , defined by @xmath162 , are identical to those for a constant potential .
thus , there are also two physically equivalent de sitter fixed points @xmath168 at @xmath169 , @xmath170 . as for the previous @xmath1-models , we use the regularity of the equations to include the above boundary subset in our analysis . by solving the constraint using eq .
, we obtain the following dynamical system : [ dynscalar5 ] @xmath171 changing @xmath63 to @xmath77 and the independent variable @xmath154 to @xmath172 according to @xmath173 results in [ dynscalar6 ] @xmath174 apart from including the past boundary , which in the present variables is given by @xmath175 , we also include the future boundary @xmath114 , which corresponds to @xmath115 and the final minkowski state . the resulting extended state space is therefore given by a finite cylinder with the region @xmath176 removed ( see fig .
[ fig : tstatespace ] ) .
-models with @xmath177 .
, scaledwidth=45.0% ] as for the @xmath1-models , since @xmath22 is monotonically decreasing , it follows that @xmath63 and @xmath77 are monotonically increasing .
again , further insight is gained by considering how @xmath63 , and hence @xmath77 , behave when @xmath178 , where @xmath119 : @xmath179 since @xmath121 when @xmath119 , it follows that @xmath63 , and hence @xmath77 , is monotonically increasing , although the graphs of @xmath63 and @xmath77 go through inflection points when @xmath119 . furthermore , since @xmath180 it follows that @xmath69 is monotonically decreasing at @xmath119 , and thus that the solution curves in the @xmath125 state space also for the @xmath2 models become horizontal in @xmath77 at @xmath119 ( see fig .
[ fig : tsolspace ] ) . from the monotonicity of @xmath77 , and the discrete symmetry that makes the two fixed points @xmath137 and @xmath138 physically equivalent , it follows that both these fixed points are sources , corresponding to asymptotic massless self - similar states .
the two physically equivalent fixed points @xmath181 and @xmath182 are center saddles , each yielding a single ( physically equivalent ) `` inflationary attractor solution '' entering the physical state space , which , as for the @xmath1-models , corresponds to a center manifold .
a center manifold expansion gives the following approximation for the inflationary attractor solution ( without loss of generality , we choose @xmath181 and @xmath183 ) : @xmath184 as in the @xmath1-model case , the periodic orbit @xmath127 corresponds to a blowup of the degenerate minkowski fixed point in the @xmath43 formulation , where @xmath127 is the future attractor .
the overall global solution structure is depicted in fig .
[ fig : tsolspace ] .
all physical solutions originate from @xmath137 and @xmath138 ( forming two physically equivalent sets of solutions ) , apart from the two physically equivalent inflationary attractor solutions that originate from @xmath168 , and all solutions end at the minkowski state associated with the future attractor and limit cycle @xmath127 . in this case
it is quite easy to translate the results to the original scalar field picture .
a similar analysis to that for the @xmath1-models shows that the open sets of physically equivalent solutions that originate from the massless states @xmath138 and
@xmath137 correspond to initial states for which @xmath148 and @xmath39 , respectively .
the single solutions that come from the de sitter fixed points @xmath181 and @xmath182 correspond to the limits @xmath35 and @xmath39 , respectively , from which they slowly evolve . as in the @xmath1-model case ,
all solutions end at the minkowski state associated with the limit cycle @xmath127 .
we begin this final section with some remarks on the relationship between the center manifold and the slow - roll approximation for the inflationary attractor solution . in the slow - roll approximation @xmath185 ( i.e. @xmath186 )
is inserted into @xmath187 , which gives @xmath188 in terms of @xmath189 , this yields the following expressions for @xmath1- and @xmath2-models : in the vicinity of the asymptotic de sitter state , where @xmath191 and @xmath192 , and therefore @xmath193 [ recall that @xmath194 , @xmath195 , these expressions yield @xmath196 to lowest order .
it follows that the slow - roll approximation leads to a curve in the @xmath75 state space that is tangential to the center manifold in the limit toward the de sitter state from which the center manifold , i.e. the inflationary attractor solution , originates , as is also true for monomial potentials as discussed in @xcite .
the inflationary `` attractor '' solution , being a one - dimensional center manifold , attracts nearby solutions exponentially rapidly , which then move along the center manifold in a relatively slow power - law manner in the vicinity of the de sitter fixed point .
thus , the center manifold structure explains both the attracting nature of the inflationary `` attractor '' solution and the fact that nearby solutions obtain a sufficient number of @xmath3-folds to be physically viable in an inflationary context .
nevertheless , although this holds for an open set of solutions that shadow the past boundary from fixed points that are sources to a de sitter state on this boundary , it should also be pointed out that there exists an open set of solutions that behave differently , as seen in figs .
[ fig : esolspace ] and [ fig : tsolspace ] . ruling out these other solutions as physically irrelevant and explaining the special role of the `` inflationary attractor solution '' beyond its center manifold structure , thereby relies on paradigmatic assumptions relating the problem to broader contexts .
examples of such contexts involve various proposed theoretical frameworks as well as , e.g. , scale considerations , illustrated in the discussion of e.g. ref .
@xcite , and various measures , motivated by , e.g. , symplectic structures ; for a recent discussion on measures which might be applicable to the present models , see @xcite .
is funded by the fct grant no .
sfrh / bpd/85194/2012 , and supported by project no .
ptdc / mat - ana/1275/2014 , and camgsd , instituto superior tcnico by fct / portugal through uid / mat/04459/2013 .
c.u . would like to thank the camgsd , instituto superior tcnico in lisbon for kind hospitality .
r. kallosh and a. linde , universality class in conformal inflation , j. cosmol . and
astropart . phys .
07 ( 2013 ) 002 . s. ferrara , r. kallosh , a. linde , and m. porrati , minimal supergravity models of inflation , phys .
d * 88 * , 085038 ( 2013 ) .
r. kallosh , a. linde , and d. roest , superconformal inflationary @xmath0-attractors , j. high energy phys .
* 11 * ( 2013 ) 198 .
r. kallosh , a. linde , and d. roest , large field inflation and double @xmath0-attractors , j. high energy phys . *
08 * ( 2014 ) 052 .
m. galante , r. kallosh , a. linde , and d. roest , unity of cosmological inflation attractors , phys .
lett . * 114 * , 141302 ( 2015 ) .
j. j. m. carrasco , r. kallosh , and a. linde , cosmological attractors and initial conditions for inflation , phys .
d * 92 * , 063519 ( 2015 ) . j. j. m. carrasco , r. kallosh , and a. linde , @xmath0-attractors : planck , lhc and dark energy , j. high energy phys . 10 ( 2015 ) 147 .
r. kallosh , and a. linde , planck , lhc , and @xmath0-attractors , phys .
d * 91 * 083528 ( 2015 ) .
a. linde , single - field @xmath0-attractors , j. cosmol .
astropart . phys .
02 ( 2015 ) 003 .
a. linde , gravitational waves and large field inflation , j. cosmol . astropart . phys .
02 ( 2017 ) 006 .
w. c. lim , h. van elst , c. uggla , and j. wainwright , asymptotic isotropization in inhomogeneous cosmology , phys .
d * 69 * 103507 ( 2004 ) . c. uggla , recent developments concerning generic spacelike singularities , gen .
bf 45 1669 ( 2013 ) .
a. alho and c. uggla , global dynamics and inflationary center manifold and slow - roll approximants , j. math .
phys . * 56 * , 012502 ( 2015 ) .
a. alho , j. hell and c. uggla , global dynamics and asymptotics for monomial scalar field potentials and perfect fluids , classical quantum gravity * 32 * , ( 145005 ) ( 2015 ) .
c. uggla , global cosmological dynamics for the scalar field representation of the modified chaplygin gas , phys .
d*88 * , 064040 ( 2013 ) .
a. alho and c. uggla , scalar field deformations of @xmath197cdm cosmology , phys .
rev . d*92 * , 103502 ( 2015 ) .
j. j. halliwell , scalar fields in cosmology with an exponential potential , phys .
b , * 185 * , 341 ( 1987 ) .
e. j. copeland , a. r. liddle , and d. wands , exponential potentials and cosmological scaling solutions , phys .
d*57 * , 4686 ( 1998 ) .
a. a. coley , , ( kluwer academic publishers , dordrecht , 2003 ) .
r. giambo and j. miritzis , energy exchange for homogeneous and isotropic universes with a scalar field coupled to matter , class .
* 27 * 095003 ( 2010 ) .
n. tamanini .
dynamical systems in dark energy models . ph.d .
thesis , university college , london , 2014 .
a. alho , s. carloni , and c. uggla . on dynamical systems approaches and methods in @xmath198 cosmology .
j. cosmol .
phys . 08 (
2016 ) 064 .
j. wainwright and g. f. r. ellis . .
( cambridge university press , cambridge , england , 1997 ) .
j. d. barrow and s. hervik . on the evolution of universes in quadratic theories of gravity .
d * 74 * 124017 ( 2006 ) .
r. grumitt , and d. sloan , measures in mutlifield inflation , arxiv:1609.05069 ( 2016 ) . | we study flat friedmann - lematre - robertson - walker @xmath0-attractor @xmath1- and @xmath2-models by introducing a dynamical systems framework that yields regularized unconstrained field equations on two - dimensional compact state spaces .
this results in both illustrative figures and a complete description of the entire solution spaces of these models , including asymptotics .
in particular , it is shown that observational viability , which requires a sufficient number of @xmath3-folds , is associated with a particular solution given by a one - dimensional center manifold of a past asymptotic de sitter state , where the center manifold structure also explains why nearby solutions are attracted to this `` inflationary attractor solution '' .
a center manifold expansion yields a description of the inflationary regime with arbitrary analytic accuracy , where the slow - roll approximation asymptotically describes the tangency condition of the center manifold at the asymptotic de sitter state . |
let @xmath0 be a simplicial complex with ground set @xmath1 . for @xmath7 let @xmath8 .
the _ alexander dual _ of @xmath0 is the simplicial complex on the same ground set defined by @xmath9 see figure [ figs ] for an example of a simplicial complex and its alexander dual . and its dual . ]
it is easy to see that @xmath10 .
furthermore , a close homological connection exists between @xmath0 and @xmath11 , that in the combinatorics folklore has become known as `` combinatorial alexander duality '' .
it states that knowledge of the homology of a simplicial complex gives knowledge of the cohomology of its alexander dual : [ adthm ] let @xmath0 be a simplicial complex with a ground set of the size @xmath12
. then @xmath13 ( here @xmath14 stands for reduced homology resp .
cohomology over a given ring @xmath6 . ) the earliest explicit statements of theorem [ adthm ] that we know of appear in kalai @xcite and stanley @xcite .
combinatorial alexander duality is a special case of the original alexander duality : ( alexander @xcite , 1922 ) let @xmath15 be a subset of the sphere @xmath16 , such that the pair @xmath17 is triangulable . then @xmath18 the connection is the following : suppose that @xmath0 is a simplicial complex different from the full simplex with ground set @xmath19 .
let @xmath20 be the @xmath12-skeleton of the full simplex on the set @xmath21 .
let us denote by @xmath22 ( resp .
@xmath23 ) a geometric realization of @xmath0 ( resp .
@xmath24 ) .
then @xmath25 , and it can be shown that @xmath26 is homotopy equivalent to @xmath27 . thus if @xmath28 is seen as a subset of @xmath16 we get @xmath29 the alexander duality theorem has played a very important role in the development of algebraic topology .
see @xcite or @xcite for context and modern treatments , and @xcite for interesting historical information about james waddell alexander , the man and his mathematics .
the modest task of this article is to make the ideas behind combinatorial alexander duality more widely accessible by giving a `` combinatorial '' proof from first principles , not relying on more general techniques .
central to our approach is a poset point of view and scrutiny of the combinatorics of the sign labeling of edges in the boolean lattice @xmath30 used for the ( co)boundary operations in ( co)homology . applications of combinatorial alexander duality in combinatorics and algebra can be found for example in @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
combinatorial alexander duality exists in more general versions .
one such generalization involves induced subcomplexes of @xmath0 versus links of faces of @xmath5 , see @xcite . for another one ,
let @xmath0 be a subcomplex of the boundary complex of a @xmath31-dimensional convex polytope @xmath32 and let @xmath5 be the subcomplex of the boundary of the dual polytope @xmath33 consisting of faces dual to those faces of @xmath32 that are not in @xmath0 .
then theorem [ adthm ] still holds , see e.g. @xcite for an application .
furthermore , a very general alexander duality theorem for nonacyclic gorenstein complexes appears in @xcite .
the reason that we confine this article to the particular simplicial setting of theorem [ adthm ] , although the idea of the proof is correct in greater generality , is that we want to work with the explicit sign labeling offered by the environment of the boolean lattice .
let @xmath0 be a simplicial complex with ground set @xmath34 . for @xmath35
we define the _ sign _ @xmath36 as @xmath37 , where @xmath38 is the @xmath3-th smallest element of the set @xmath39 .
the following simple property of the sign function will be needed .
[ signs ] let @xmath40 and @xmath41 then @xmath42 we have that @xmath43 and @xmath44 in the rest of this section we review the definitions and notation used for ( co)homology . throughout the paper suppose that @xmath45 is a commutative ring containing a unit element . let @xmath46 be a free @xmath6-module with the free basis @xmath47 . the _ reduced chain complex _ of @xmath0 over @xmath6 is the complex @xmath48 whose mappings @xmath49 are defined as @xmath50 the complex @xmath51 is formally infinite ; however , @xmath52 for @xmath53 or @xmath54 .
@xmath12-th reduced homology group _ of @xmath0 over @xmath6 is defined as @xmath55 let @xmath56 be a free @xmath6-module with the free basis @xmath57 .
the _ reduced cochain complex _ of @xmath0 over @xmath6 is the complex @xmath58 where @xmath59 are maps dual to @xmath49 , explicitly stated : @xmath60 the _ @xmath12-th reduced cohomology group _ of @xmath0 over @xmath6 is defined as @xmath61 suppose that @xmath0 is a simplicial complex and @xmath62 is a subcomplex of @xmath0 .
let @xmath63 , where @xmath64 was defined in section [ redhom ] .
the _ relative reduced chain complex _ of @xmath65 over @xmath6 is the complex @xmath66 where @xmath67 are defined as @xmath68 the _ @xmath12-th relative reduced homology group _ of @xmath65 over @xmath6 is defined as @xmath69 [ relhom ] when we wish to compute relative homology groups , we can identify @xmath70 with a free @xmath6-module with the free basis @xmath71 .
then @xmath67 can be rewritten as : @xmath72 one of the important properties of relative homology groups is that they fit into a long exact sequence .
see e. g. @xcite or @xcite for a proof and more details .
[ lespair ] suppose that @xmath0 and @xmath62 are simplicial complexes @xmath73 .
then there is a long exact sequence @xmath74
before embarking on the proof of theorem [ adthm ] , we first present the idea .
suppose that @xmath0 is a simplicial complex with ground set @xmath1 .
let @xmath75 be the lattice of all subsets of @xmath1 , and let @xmath76 be the subposet of @xmath75 corresponding to the subsets that are in @xmath0 .
then the @xmath12-th homology group of @xmath0 depends just on the @xmath12-th and @xmath77-st levels of the poset @xmath76 .
let @xmath30 be a full simplex with vertex set @xmath1 .
it is easy to see that @xmath78 ( see lemma [ adfirst ] ) .
thus , we restate the problem as computing homologies of the chain complex determined by the complement of @xmath76 in the sense of remark [ relhom ] .
the idea of the proof is that if we turn the lattice upside down ( exchange @xmath79 with its complement ) , then this combinatorial map on generators should induce a canonical isomorphism between the relative homology of the pair @xmath80 and the cohomology of @xmath5 .
this idea is basically correct ; however , the mapping is not an isomorphism as described some sign operations are necessary .
let @xmath81 be the simplicial complex in figure [ figs ] .
its ground set is the set @xmath82 .
the posets @xmath83 and @xmath84 are depicted in figure [ figgammas ] , and the left part of the picture also shows the complement of @xmath83 ( bold , dashed ) determining the homology of @xmath85 . in the sense of remark [ relhom ] , the chain complex @xmath86 is @xmath87 and the cochain complex @xmath88 is @xmath89 the map @xmath90 is _ not _ an isomorphism of these two complexes ( if @xmath91 for @xmath92 ) , since @xmath93 , while @xmath94 . nevertheless
, these two chain complexes are isomorphic , as will be shown in the next section .
( left ) and @xmath84 ( right ) . ]
the proof of theorem [ adthm ] is obtained by combining lemma [ adfirst ] and lemma [ adsecond ] .
define @xmath110 as in lemma [ signs ] and let @xmath111 be the isomorphisms generated by the formula @xmath112 for @xmath113 , @xmath114 ( note that these two conditions are equivalent to @xmath115 , @xmath116 ) .
we then have a diagram : we check that @xmath117 .
let @xmath118 , @xmath113 , @xmath119 . then @xmath120 @xmath121 these two sums are equal term by term , due to lemma [ signs ] .
thus @xmath122 is an isomorphism of the complexes , implying @xmath123 | let @xmath0 be a simplicial complex with ground set @xmath1 . define its alexander dual as the simplicial complex @xmath2 .
the combinatorial alexander duality states that the @xmath3-th reduced homology group of @xmath0 is isomorphic to the @xmath4-th reduced cohomology group of @xmath5 ( over a given commutative ring @xmath6 ) .
we give a self - contained proof from first principles , accessible to the non - expert . |
large scale isotropy and homogeneity is the most fundamental assumption of modern cosmology .
those assumptions make it possible to solve einstein s equations , which reduce to two differential equations with only one parameter , the scale factor . to test this hypothesis and to estimate the effect of the deviations from homogeneity and isotropy
an averaging formalism was developed in @xcite .
it has been critisized in @xcite that this approach relies on an a priory fixed frame and is thus gauge dependent .
an alternative has been suggest which is valid for infinite spacial volumes . in this work
we use a physical reference frame , the frame of comoving dust , to perform the averaging in a manifestly gauge independent way .
our formalism is applicable to finite volumes . at the end
we compare our result to the equations obtained in @xcite and show that they are equivalent in the frame of the introduced pressure - less fluid .
starting from local inhomogeneous einstein equations an averaging procedure is set up to derive effective equations which quantitatively measure the deviation from the friedmann equations which describe the dynamics of a homogeneous and isotropic space - time . to avoid the question of coordinate fixing and
hence gauge dependence a system of freely falling observers will be introduced and used to perform the averaging .
it is important to mention the compatibility of the dust as coordinate system with a deformed frw universe , as it respects the fundamental symmetries of the background .
we know that our universe used to be homogeneous to a very high precision and hence was described well by the frw universe
. however , nowadays there are large inhomogeneities locally present in the universe .
nevertheless , observations of the large scale structure suggest that the underlying geometry is not drastically different from an homogeneous and isotropic space on horizon scales .
therefore , it is sensible to use a laboratory which does not destroy the frw space - time , since it has the appropriate symmetries .
the dust we are going to use fulfils this criterion . we consider a space - time filled with matter and demand that the energy density of this matter should never vanish .
this assumption allows to avoid the problem of seemingly indeterministic behaviour of general relativity ( gr ) , known as the hole problem @xcite . demanding a non vanishing matter content we are able to give physical meaning to every point in space - time .
this is also a reasonable assumption on physical grounds , since even the voids contain some amounts of cosmic dust or radiation .
furthermore , it is speculated that there is matter present in the universe , which interacts only ( or almost only ) with gravity .
the dark matter is considered to be `` cold '' i.e. non- relativistic in the standard model of cosmology , hence it can be well described by dust .
thus , our model can be viewed as a realistic cosmological model incorporating the cold dark matter .
the distribution of matter in our model though is not homogeneous , since it is supposed to represent a universe perturbed about the frw background symmetry .
this is the difference to the approach in @xcite , we allow for matter to exist which is distributed inhomogeneously and hence it is possible to find physical fields with a space - like gradient . within our model
the energy momentum tensor can be decomposed in the way described below , since we have demanded for the energy density never to vanish .
we can view this in two ways .
either the dust of low energy is formally separated from the total energy momentum tensor , or it is added to it , which would not affect the system strongly .
@xmath0 where the @xmath1 is the dust energy momentum with infinitesimal @xmath2 and @xmath3 is the energy momentum tensor of the rest of the system .
consequently the lagrangian of the theory is as follows @xmath4 the gravitational lagrangian is the ricci scalar multiplied by the square root of the metric determinant .
the dust lagrangian is @xmath5 with the following eight fields , @xmath2 being the energy density , @xmath6 eigentime , @xmath7 velocity field and @xmath8 coordinate field , where @xmath9 @xmath10 @xmath6 has a time - like gradient and the @xmath8 s have obviously space - like gradients .
this fields will be used for the construction of a window function which transforms as a scalar .
we will restrict our analysis to the lagrangian formalism and not be concerned with the form of the physical hamiltonian . on the other hand we know that a physical hamiltonian exists , since the system with dust deparametrizes , as has been shown in @xcite .
the eigentime of the dust particles is a map into the dust - time manifold @xmath11 at each fixed value @xmath12 , @xmath6 is a scalar w.r.t .
coordinate changes on the manifold @xmath13 in the sense @xmath14 where @xmath15 . on the dust - time manifold @xmath6
is trivially a scalar
. this will be scalar with a space - like gradient which will be used for the construction of the window function .
we take the eigentime @xmath6 of the dust particles as the scalar field which generates the time flow .
therefore the unit time - like vector is @xmath16 the foliation is defined as the hypersurfaces on which @xmath6 takes constant values .
it is the physical time of the dust field .
this defines a time - space split with the projector orthogonal to the time flow @xmath17 the dust is a collection of geodesic observers , freely falling on @xmath13 .
each particle carries a label , at any instant of time @xmath18 a particle has a position , so this position of the labeled particle is a physical value and hence a dirac observable . taking the continuum limit ,
while keeping the particle density constant we obtain a field , where the indices of the particles can be viewed now as three dimensional coordinates @xmath19 .
@xmath20 is a map from the hypersurface of constant time into the dust - space manifold of constant @xmath12 @xmath21 now we define the quantity : @xmath22 .
the @xmath23 are for evry fixed pair of values @xmath24 and @xmath6 scalars on @xmath13 , ( omit the bar in the coming discussion ) @xmath25 where @xmath15 .
on the other hand they are three vectors on @xmath26 .
to construct a quantity which is a scalar on @xmath13 and @xmath26 , more work is required .
the basic idea is to construct a scalar on @xmath26 out of the coordinate fields of the dust , a quantity like @xmath27 .
therefore we need the induced metric on the hypersurfaces @xmath28 . to get this object
we pull back the orthogonal projector @xmath29 on the hypersurface of constant @xmath6 denoted by @xmath30 .
to stay general we introduce coordinates on @xmath30 called @xmath31 which represent a map @xmath32 furthermore , the coordinates on @xmath13 are denoted by @xmath33 and represent a map @xmath34 the combined map which we are going to use for the pull - back is @xmath35 , defined as @xmath36 we can visualize the interrelations as follows , the hypersurfaces of constant dust - time are denoted by @xmath30 and @xmath37 .
while the map @xmath38 goes from the dust - space to the space of internal coordinates of the hypersurface of constant dust - time , the map @xmath39 goes from this internal space to the manifold .
therefore , the combination of both maps goes from the dust - space slice to the full space - time manifold and is used for the pull - back . ] with the pull backs of the maps given as @xmath40 @xmath41 @xmath42 the choice of the @xmath39 and @xmath38 maps corresponds to the gauge choice .
later we will evaluate the result in the the so called adm gauge where @xmath43 and @xmath44 , therefore @xmath45 .
after defining the maps we come back to the main goal and pull - back on the dust space @xmath46 here the property has been used that @xmath47 is orthogonal to the tangents of @xmath26 : @xmath48 , i.e. @xmath49 . note : for the construction we have used the orthogonality of the time flow to the spatial hypersurfaces . to ensure this
we can pick the dust velocity field @xmath50 to vanish .
this is simply assuming that the dust we consider is rotation free , which is reasonable for all practical purposes . the constructed quantity commutes with the 3-diffeomorphism constraints and in the next step we construct the observable , using the dynamic equation with @xmath51 being @xmath52 from the adm split : @xmath53 here @xmath54 with @xmath55 some scalar density and @xmath56 the total hamiltonian constraint of the system .
even if an analytic solution to this equation can not be obtained , we can set initial values by field re - definition and perform a series expansion .
the solution of this equation @xmath57 is used to define the observable @xmath58 @xmath59 this quantity commutes for fixed @xmath6 and @xmath20 with all constraints and is hence an observable . in the following
we will omit the bar for simplicity .
now we found the tensor quantity on @xmath26 which is needed for the contraction with the vectors and can write @xmath60 the dust vectors are contracted with a tensor on the dust manifold and one obtains a physical scalar on @xmath13 and on @xmath26 .
it transforms as @xmath61 and @xmath62 where on @xmath13 : @xmath15 and correspondingly on @xmath26 : @xmath63 .
we check the transformation properties below . * the induced metric on the hypersurface @xmath64 is a scalar under coordinate changes on the manifold by construction , an explicit calculation shows it in components : + @xmath65 * the constructed quantity @xmath66 is a scalar on @xmath26 since the 3-vectors @xmath8
are contracted with a rank two tensor on @xmath26 .
again explicit calculation shows : @xmath67 * the connecting relation which makes the whole object invariant under coordinate transformations on the manifold is that the @xmath8 s are the dust labels and hence scalars on @xmath13 as discussed above : + @xmath25 + where @xmath15 .
now we have shown that @xmath66 is a scalar on @xmath13 and on the dust manifold @xmath26 . since the @xmath68 is a scalar on @xmath26 we can express it in arbitrary coordinates @xmath31 on @xmath30 , the adm split hypersurface of constant @xmath6 .
this coordinate representation is also convenient for the window function which will be discussed in the following section .
having constructed physical scalars we write the window function as @xmath69 with @xmath70 is a dirac delta function and @xmath71 the step function .
the window function constructed in ( [ eqn : window ] ) is a scalar on @xmath13 and @xmath26 i.e. it transforms as @xmath72 now the functional is set up in the following way @xmath73 the averaging functional of a quantity @xmath26 can be defined using eqn .
( [ eqn : functional ] ) , and simplified integrating out the delta function leading to @xmath74 where @xmath75 is the hypersurface on which @xmath76 , @xmath77 the determinant of the induced spatial metric discussed above and @xmath78 the lapse function for a general not synchronous gauge .
the average is manifestly a gauge independent quantity , since @xmath79 and @xmath80 are scalars under diff(@xmath81 ) and also diff(@xmath26 ) and so the window function is also a scalar , leading to the following transformation property of the averaging functional @xmath82 introducing @xmath83 we obtain @xmath84 the gauge functional is obviously gauge invariant and hence the average of @xmath26 is an observable .
equipped with this functional we can address the problem of averaging the scalar parts of einstein s equations of the model of the universe presented above , where the coordinate system , the dust , is included into the studied system .
we will perform a similar analysis as in @xcite , but in a more general way , since our formalism applies also to finite volumes . to average the dynamic equations of our model it is also necessary to compute time derivatives of the averaged quantities . in this case the time derivative means partial derivative w.r.t @xmath85 the eigentime of the homogeneous dust used for the deparametrisation .
let us begin with the derivative of @xmath86
. the strategy will be to choose certain coordinates in which the computation will be easier and then to generalize our result to an arbitrary frame .
the eigentime @xmath6 is the function that defines the space as hypersurfaces on which it assumes a constant value , therefore , it is spatially homogeneous . in this case
we can choose coordinates , in future referred to as adm coordinates , such that : @xmath87 and @xmath88 where @xmath78 is the lapse and @xmath89 the shift .
the metric reads : @xmath90 furthermore , since we have set the velocity field @xmath91 of the coordinate dust to zero , which implies the orthogonality of @xmath47 to the spatial hypersurfaces , the shift vectors vanish . in this coordinates we compute the time derivative .
the partial derivative commutes past the integral and the derivative of the @xmath92 distribution we understand in the weak sense @xmath93 in the coordinates chosen we can read of from the form of @xmath94 the only component of @xmath95 which is non - zero : @xmath96 and @xmath97 with @xmath98 .
thus , eqn .
( [ eqn : derivativeaverage ] ) simplifies to @xmath99 here @xmath100 and @xmath101 , since @xmath102 in this gauge and the trace @xmath103 . having this
, we can rewrite the expression in a covariant way , so that it reduces to our result when the adm coordinates are fixed @xmath104 the last term vanishes in case @xmath105 i.e. the spatial coordinates do not depend on the time variable .
this is indeed the case in our model since @xmath47 is orthogonal to the space such that @xmath106 which implies the above relation .
hence the time derivative reads @xmath107 now we can calculate the time derivative of the average of a scalar @xmath108 covariantly expressed we obtain the general form of the buchert , ehlers commutation rule , @xcite : @xmath109 in order to be interpreted as a domain dependent scale factor analogous to the @xmath110 of the frw model , a quantity has to be considered which is gauge independent . as any physical observable must not depend on the coordinates chosen . to accomplish this we use the gauge independent averaging functional to define the quantity @xmath111 to be the effective scale factor @xmath112 using ( [ eq : derivative of functional ] )
one sees that @xmath113 and hence @xmath114 a more pictorial way to describe this , is to write @xmath115 leading to @xmath116 we see from this calculation that @xmath111 has indeed a property of a scale factor , but its definition above circumvent the necessity of introducing a normalization at a certain initial time point .
we will derive the expressions for the averaged hamilton constraint density and the raychaudhuri equation , the important point of the derivations is that in our set up the @xmath117 , which defines the time flow of the coordinate dust , is already present in the model and is not chosen arbitrarily . the time arises natural and there is no need of artificial gauge fixing .
thus , we describe the time flow without formally choosing a gauge .
the einstein tensor multiplied with the time flow vectors ( which is equivalent to the hamiltonian constraint density ) gives @xmath118 where the expansion rate and the shear are defined w.r.t @xmath94 , the @xmath119 is the spatial curvature associated with the projector on the hypersurface orthogonal to @xmath94 . on the other hand one has from the matter contribution
@xmath120 with @xmath121 since we said that the velocity field @xmath91 vanishes .
now the hamiltonian constraint density can be averaged . for normalization
divide it first by @xmath122 apply the averaging functional introduced above and insert the identity by adding and subtracting the average of @xmath123 squared we get @xmath124 using @xmath125 and defining the backreaction @xmath126 we rewrite ( [ eq : hamilton constraint ] ) as @xmath127 this is the effective equation describing the velocity of the expansion of the domain dependent scale factor .
it is the analogue to the first friedmann equation . with the definitions comparable to those in @xcite we can write the cosmological energy balance : @xmath128 @xmath129 @xmath130 @xmath131 @xmath132 @xmath133 from the energy balance equation we see that the backreaction term @xmath134 as well as the coordinate mass term
@xmath135 enter the balance and the backreaction term opens the possibility of contributing to an effective cosmological constant term . combining the hamiltonian constraint density with the trace of the einstein equations projected orthogonal to the time flow
, one gets @xmath136 which is equivalent to @xmath137 in our model the last term gives @xmath138 now we will need the second time derivative of the scale factor s : @xmath139 we will use the commutator ( [ eq : commutation rule ] ) and eqn . ( [ eq : adot ] ) to get @xmath140 rewriting @xmath141 we can insert ( [ eq : rauchaudhuri ] ) into the first term @xmath142 this is the effective raychaudhuri equation for the domain .
there is an analogy to the second friedman equation , since it describes the acceleration of the domain dependent scale factor s. in this subsection we will evaluate the general equations in the adm gauge , which was introduced above , and compare them to the effective equations obtained by t. buchert . as a reminder in the adm gauge @xmath87 and @xmath88 where @xmath78 is the lapse and @xmath89 the shift .
since we have chosen @xmath94 to be orthogonal to the spatial hypersurfaces the shift vectors vanish i.e. the metric reads : @xmath143 .
especially we get for @xmath144 and @xmath145 since @xmath85 defines the hypersurface of constant physical time .
given that , the hamiltonian constraint reduces to @xmath146 and the raychaudhuri equation reads as @xmath147 setting @xmath148 we obtain : the hamilton constraint density @xmath149 and the raychaudhuri equation : @xmath150 note : the coordinate dust energy @xmath2 has a negative effect on the acceleration .
the averaging functional with the window function reduces in this gauge to @xmath151 to perform integration in general we need a push forward from the manifold into the @xmath152 , in our case a map from the dust space into @xmath152 is required .
define : @xmath153 s.t .
@xmath154 where @xmath155 is the map @xmath156 , @xmath38 the map @xmath157 , @xmath39 the map @xmath158 and @xmath159 the dust - space slice . used for the push forward into @xmath152 .
this map is a combination of maps from the dust - space slice of constant dust - time to the space of internal variables : @xmath38 , the map from this internal space to the manifold @xmath81 : @xmath39 and the map form the manifold to the @xmath152 : @xmath155 . ]
it was already shown that @xmath68 is gauge invariant so now it is legitimate to fix a gauge .
a legitimate choice would be @xmath160 the identity , @xmath39 the map from the hypersurface coordinates @xmath161 to the manifold coordinates @xmath162 and therefore @xmath163 . in the adm gauge the coordinates are @xmath51 and @xmath31 as read of from the metric @xmath164 .
this leads to @xmath165 rewriting the average functional one obtains @xmath166 which is a volume integral over a sphere of radius @xmath167 . with this formalism
the effect of inhomogeneity on an effective friedmann model can be estimated and an example of application to the ideal fluid cosmology can be found in the appendix [ app : ifc ] .
the difference between our formalism and the averaging applied by t. buchert in @xcite is that we did not choose a gauge from the beginning but deparametrized the manifold with dust .
this introduced a time in the theory .
furthermore we found a way to perform gauge invariant averages over finite volumes .
hence the averaged quantities @xmath168 became physical observables and we obtained equations governing their dynamics .
the equations we get after fixing the adm gauge are slightly different to the ones in @xcite .
we recover buchert s equations in our general formalism if we set the rest energy momentum @xmath169 to zero .
therefore one could think that in buchert s formalism just the geometry is averaged .
the problem with this point of view is that even an infinitesimal energy density @xmath2 leads to a geometry arbitrary far away from the vacuum geometry , since there is no smooth limit @xmath170 .
hence one has to see that buchert s equations describe correctly the evolution of a cosmology filled with one type of pressure - free fluid in the rest frame of this fluid .
we see that in this special case the procedures of gauge fixing and averaging commute .
our formalism allows to generalize the averaging to an almost arbitrary energy momentum tensor .
the only restriction is that the energy density must not vanish anywhere .
especially it is not possible to apply this formalism to space - time devoid of all matter .
a manifestly gauge invariant averaging formalism for finite volumes was presented in this work .
the question of gauge dependence of the backreaction was resolved .
the model used just required the assumption of a non - vanishing energy density . in this case
a dust energy momentum tensor can be separated from the total energy density .
the coordinate fields of the dust were used as a reference frame to perform averaging .
the derived equations for the average of an arbitrary energy momentum tensor were evaluated in the adm gauge , which corresponds to choosing the rest frame of the dust .
furthermore , as an example the equation is evaluated for an energy momentum tensor of an ideal fluid .
the resulting equation was also studied in the adm gauge .
an interesting phenomenon was observed , as discussed in the ideal fluid cosmology in the appendix [ app : ifc ] .
it appears that `` tilt '' effects ( i.e. non - co - linearity of the dust flow and the fluid s flow ) contribute negatively to the acceleration of the domain dependent scale factor .
the derived equations in the adm gauge were compared to the buchert equations .
we are able to interpret the buchert equations in the gauge independent framework .
it turned out that they describe the averaged behaviour of a pressure - free fluid , using this fluid itself as a reference and choosing it s rest frame .
it turns out that in this special case gauge fixing and averaging commute . by deriving gauge independent equations for the cosmological backreaction
, we proved that it is in principle an observable . with this formalism at hand
it is possible to study cosmological data and can trust the derived equations to correctly describe the observations , once we identify a system of freely falling observers .
one could go one step further and argue that given the necessity to explain the abundance of a matter species , which mainly -or entirely- interacts gravitationally , we can view our system with the coordinate dust as a realistic cosmological model . in which case the dust is the dark matter which behaves , according to observations as a pressure - less fluid .
an interesting and may be fundamental observation is , that it is not possible to define averaged observables in a space - time devoid of matter .
i would like to thank k. giesel , s. meyer and m. bartelmann for very helpful discussions .
i acknowledge support from the imprs for precision tests of fundamental symmetries .
most warmly i would like to thank s. hofmann who has suggested to study this topic and has provided vital guidance during the completion of this work .
a cosmology equipped with an energy momentum tensor of an ideal fluid will be studied .
this is a reasonable choice for the energy momentum , since it is the most general tensor fulfilling the symmetries of an isotropic universe and since we know from the cmb that the universe has been extremely isotropic this model is a reasonable choice
. this procedure can be easily generalized to an arbitrary number of ideal non interacting fluids . where the expansion rate and the shear are defined w.r.t @xmath94
, the @xmath119 is the spatial curvature associated with the projector on the hypersurface orthogonal to @xmath94 . on the other hand one has from the matter contribution where the tilt angle @xmath175 defined via @xmath176 .
furthermore as discussed above @xmath121 since we said that the velocity field @xmath91 vanishes .
now the hamiltonian constraint can be averaged . for normalization
divide it first by @xmath122 and apply the averaging functional introduced above from the energy balance equation we see that the backreaction term @xmath134 opens the possibility of contributing to an effective cosmological constant term in this model , which is in principle close to @xmath180-cdm . *
both the hamiltonian constraint and the raychaudhuri equation contain the energy density of the coordinate dust . in the second equation
it contributes negatively to the acceleration of the scale factor .
* the fact that generically the flow of the coordinate dust is non parallel to the flow of the fluid which is averaged , is manifested in the @xmath190 term .
this as well contributes negatively to the acceleration of the scale factor @xmath111 .
* after the gauge choice the spatial integral is an ordinary 3-dimensional integral . *
the backreaction @xmath191 has the potential to contribute to the acceleration .
this has been addressed in the literature @xcite and certainly deserves attention . | for the study of cosmological backreacktion an avaragng procedure is required . in this work a covariant and gauge invariant averaging formalism for finite volumes will be developed .
this averaging will be applied to the scalar parts of einstein s equations . for this purpose `` dust '' as a physical laboratory
will be coupled to the gravitating system .
the goal is to study the deviation from the homogeneous universe and the impact of this deviation on the dynamics of our universe .
fields of physical observers are included in the studied system and used to construct a reference frame to perform the averaging without a formal gauge fixing .
the derived equations resolve the question whether backreaction is gauge dependent . |
a _ market _ is a set of arrangements by which buyers and sellers , collectively known as _ traders _ , are in contact to exchange goods or services .
_ auctions _ , a subclass of markets with strict regulations governing the information available to traders in the market and the possible actions they can take , have been widely used in solving real - world optimization problems , and in structuring stock or futures exchanges .
the most common kind of auction is the _ english auction _ , in which there is a single seller , and multiple buyers compete by making increasing bids for the commodity ( good or service ) being auctioned ; the one who offers the highest price wins the right to purchase the commodity . since only one type of trader buyers
makes offers in an english auction , the auction belongs to the class of _ single - sided auctions_. another common single - sided auction is the _ dutch auction _ , in which the auctioneer initially calls out a high price and then gradually lowers it until one bidder indicates they will accept that price .
another class of single - sided auctions is the class of _ sealed - bid auctions _ , in which all buyers submit a single bid and do so simultaneously , i.e. , without observing the bids of the others or if the others have bid .
two common sealed - bid auctions are the _ first - price auction _ and the _ second - price auction _ or _ vickrey auction _ @xcite . in both types of sealed - bid auctions ,
the highest bidder obtains the commodity . in the former
, the highest bidder pays the price they bid , while in the latter , they pay the second highest price that was bid . these four single - sided auctions
english , dutch , first - price sealed - bid , and vickrey are commonly referred to as the _ standard auctions _ and were the basis of much early research on auctions .
in addition , there are _ double - sided auctions _ or s , in which both sellers and buyers make offers , or _
shouts_. the two most common forms of da are _ clearing houses _ or chs and _ continuous double auctions _ or cdas . in a ch , an auctioneer first collects _
bids_shouts from buyers and _ asks_shouts from sellers , and then clears the market at a price where the quantity of the commodity supplied equals the quantity demanded . this type of market clearing guarantees that if a given trader is involved in a transaction , all traders with more competitive offers are also involved . in a cda ,
a trader can make a shout and accept an offer from someone at any time .
this design makes a cda able to process many transactions in a short time , but permits extra - marginal traders to make deals .
both kinds of are of practical importance , with , for example , variants being widely used in real - world stock or trading markets including the new york stock exchange ( ) and the chicago mercantile exchange ( ) . in some auctions
, traders can place shouts on combinations of items , or packages " , rather than just individual items .
they are called _
combinatorial auctions_. a common procedure in these markets is to auction the individual items and then at the end to accept bids for packages of items .
combinatorial auctions present a host of new challenges as compared to traditional auctions , including the so - called _ winner determination problem _
, which is how to efficiently determine the allocation once the bids have been submitted to the auctioneer .
traders , in some cases , are allowed to both sell and buy during an auction .
such traders are called _ two - way traders _ , while those that only buy or only sell are called _ one - way traders_. this report will mainly discuss non - combinatorial s , especially s , populated by one - way traders .
a central concern in studies of auction mechanisms are the supply and demand schedules in a market .
the quantity of a commodity that buyers are prepared to purchase at each possible price is referred to as the _ demand _ , and the quantity of a commodity that sellers are prepared to sell at each possible price is referred to as the _
supply_. thus if @xmath0 is plotted as a function of @xmath1 , the _ demand curve _ slopes downward and the _ supply curve _ slopes upward , as shown in figure [ fig : underlying - supply - demand ] , since the greater the price of a commodity , the more sellers are inclined to sell and the fewer buyers are willing to buy .
typically , there is some price at which the quantity demanded is equal to the quantity supplied . graphically , this is the intersection of the demand and supply curves .
the price is called the _ equilibrium price _ , and the corresponding quantity of commodity that is traded is called the _
equilibrium quantity_. the equilibrium price and equilibrium quantity are denoted as @xmath2 and @xmath3 respectively in figure [ fig : underlying - supply - demand ] .
each trader in an auction presumably has a limit price , called its _ private value _ , below which sellers will not sell and above which buyers will not buy .
the private values of traders are not publicly known in most practical scenarios .
what is known instead are the prices that traders offer .
self - interested sellers will presumably offer higher prices than their private values to make a profit and self - interested buyers tend to offer lower prices than their private values to save money .
the prices and quantities that are offered also make a set of supply and demand curves , called the _ apparent supply and demand curves _
, while the curves based on traders private values are called the _ underlying supply and demand_. figure [ fig : apparent - supply - demand ] shows that the apparent supply curve shifts up compared to the underlying supply curve in figure [ fig : underlying - supply - demand ] , while the apparent demand curve shifts down .
when traders are excessively greedy , the apparent supply and demand curves do not intersect and thus no transactions can be made between sellers and buyers unless they compromise on their profit levels and adjust their offered prices . in a , buyers and sellers not only ` haggle ' on prices in a collective manner , but they also face competition from opponents on the same side of the market . thus buyers , for example , are not only collectively trying to drive prices down , against the wishes of sellers , but they are also individually trying to ensure that they , rather than other buyers , make profitable trades .
this leads to shouts becoming more and more competitive over time in a given market .
figure [ fig : time - series ] shows a typical time series of shouts in a .
ask prices usually start high while bid prices start low .
gradually , traders adjust their offered prices , or make new shouts , closing the gap between standing asks and bids until the price of a bid surpasses that of an ask .
such an overlap results in a transaction , shown as a solid bar between the matched ask and bid in figure [ fig : time - series ] . in the market
depicted in figure [ fig : time - series ] , newly placed bids ( asks ) do not have to beat the outstanding bids ( asks ) .
however in some variants of the including the market operated by the , new shouts must improve on existing ones .
this requirement is commonly referred to as the _ shout improvement rule _ @xcite . in some real - world stock markets , including the and the markets , trades are made through _ specialists _ or _ market makers _ , who buy or sell stock from their own inventory to keep the market liquid or to prevent rapid price changes .
each specialist is required to publish on a regular and continuous basis both a _ bid quote _ , the highest price it will pay a trader to purchase securities , and an _ ask quote _ , the lowest price it will accept from a trader to sell securities . the specialist is obligated to stand ready to buy at the bid quote or sell at the ask quote up to a certain number of shares .
the range between the lower bid quote and the higher ask quote is called the _ bid - ask spread _ , which , according to stock exchange regulations , must be suitably small . if buy orders temporarily outpace sell orders , or conversely if sell orders outpace buy orders , the specialist is required to use its own capital to minimize the imbalance .
this is done by buying or selling against the trend of the market until a price is reached at which public supply and demand are once again in balance .
maintaining a bid - ask spread creates risk for a specialist , but when well maintained , also brings huge profits , especially in an active market @xcite .
markets involving specialists that present quotes are called _ quote - driven markets_. another class of markets are _ order - driven markets _ , in which all of the orders of buyers and sellers are displayed .
this contrasts with quote - driven markets where only the orders of market makers are shown .
an example of an order - driven market is the market formed by _ electronic communication networks _ or s. these are electronic systems connecting individual traders so that they can trade directly between themselves without having to go through a middleman like a market maker .
the biggest advantage of this market type is its transparency .
the drawback is that in an order - driven market , there is no guarantee of order execution , meaning that a trader has no guarantee of making a trade at a given price , while it is guaranteed in a quote - driven market .
there are markets that combine attributes from quote- and order - driven markets to form hybrid systems
. our discussion above may give the impression that in real markets trade orders are made directly by the individuals who want to buy or sell stock . in practice , traders commonly place orders through brokerage firms , which then manage the process of executing the orders through a market .
auctions with different rules and populated by different sets of traders may vary greatly in performance .
popular performance measurements include , but are not limited to , _ allocative efficiency _ and the _ coefficient of convergence_. the allocative efficiency of an auction , denoted as @xmath4 , is used to measure how much social welfare is obtained through the auction .
the _ actual overall profit _ , @xmath5 , of an auction is : @xmath6 where @xmath7 is the transaction price of a trade completed by agent @xmath8 and @xmath9 is the private value of agent @xmath8 , where @xmath8 ranges over all agents who trade .
the _ theoretical _ or _ equilibrium profit _
, @xmath10 , of an auction is : @xmath11 for all agents whose private value is no less competitive than the equilibrium price , where @xmath12 is the equilibrium price .
given these : @xmath13 @xmath4 is thus a measure of the proportion of the theoretical profit that is achieved in practice .
the convergence coefficient , denoted as @xmath14 , was introduced by smith @xcite to measure how far an active auction is away from the equilibrium point .
it actually measures the relative deviation of transaction prices from the equilibrium price : @xmath15 since markets with human traders often trade close to the equilibrium price , @xmath14 is used as a way of telling how closely artificial traders approach human trading performance .
research on auctions originally interested mathematical economists .
they view auctions as games and have successfully applied traditional analytic methods from game theory .
this section therefore takes an overlook at basic concepts in game theory .
the games studied by game theory are well - defined mathematical objects .
a game is usually represented in its _ normal form _ , or _ strategic form _ , which is a tuple @xmath16 @xmath17 is the number of players , @xmath18 is the set of _ actions _ available to player @xmath8 , and @xmath19 is the _ payoff _ or _ utility function _
@xmath20 , where @xmath21 is the joint action space @xmath22 .
when a player needs to act , it may follow a _ pure strategy _
, choosing an action , @xmath23 , from its action set , or a _ mixed strategy _ , @xmath24 , choosing actions according to a probability distribution .
the strategy set of player @xmath8 , denoted as @xmath25 , is the same thing as a set of probability distributions over @xmath18 , denoted as @xmath26 .
a joint strategy for all players is called a _ strategy profile
_ , denoted as @xmath27 , and @xmath28 is the probability all players choose the joint action @xmath29 from @xmath21 .
thus player @xmath8 s payoff for the strategy profile @xmath27 is : @xmath30 in addition , @xmath31 denotes the set of all possible strategy profiles , . ]
@xmath32 is a strategy profile for all players except @xmath8 , and @xmath33 is the strategy profile where player @xmath8 uses strategy @xmath24 and the others use @xmath32 .
a normal - form game is typically illustrated as a matrix with each dimension listing the choices of one player and each cell containing the payoffs of players for the corresponding joint action .
figure [ fig : normal - form - prisoner - dilemma ] shows the normal form of the well - known prisoner s dilemma game .
alternatively , games may be represented in _ extensive form _ , which is a tree , as in figure [ fig : extensive - form - game ] .
the tree starts with an initial node and each node represents a state during play . at each non - terminal node
, a given player has the choice of action .
different choices lead to different child nodes , until a terminal node is reached where the game is complete and the payoffs to players are given .
@xmath34 = [ level distance=2 cm , sibling distance=2 cm ] = [ level distance=2 cm , sibling distance=1 cm ] = [ fill , circle , minimum size=1pt ] = [ right=5pt ] = [ text centered ] child node[tree ] child node[tree ] node[terminal ] @xmath35 edge from parent node[annotation , below ] @xmath36 child node[tree ] node[terminal ] @xmath37 edge from parent node[annotation , above ] @xmath38 edge from parent node[annotation , below ] @xmath39 child node[tree ] child node[tree ] node[terminal ] @xmath40 edge from parent node[annotation , below ] @xmath36 child node[tree ] node[terminal ] @xmath41 edge from parent node[annotation , above ] @xmath38 edge from parent node[annotation , above ] @xmath42 ; a game may be _ cooperative _ or _ noncooperative _ , as players in these games are respectively _ cooperative _ or _ self - interested_. cooperative players share a common payoff function , i.e. , @xmath43 whereas self - interested players typically have distinct payoff functions . in both cases ,
players need to coordinate in a certain way to ` assist ' each other in achieving their goals @xcite . if the payoffs of all players for each strategy profile sum to zero , the noncooperative game is called a _
zero - sum _ game , i.e. , @xmath44 zero - sum games are a special case of a more general class of games called _ constant - sum _ games , where the sum of all payoffs for each outcome is a constant but may not necessarily be zero .
non - zero - sum games are sometimes referred to as _ general - sum _ games . in economic situations ,
the exchange of commodities is considered general - sum , since both parties gain more through the transaction than if they had not transacted ( otherwise the exchange would not have happened , assuming both are rational ) . in some games , the payoffs for playing a particular strategy remain unchanged as long as the other strategies employed collectively by the players are same , no matter which player takes which action .
these games are called _ symmetric games _ and the rest are _
asymmetric games_. for example , the prisoner s dilemma game given above is symmetric
. players may take actions _
simultaneously _ or _
sequentially_. in a sequential game , players have alternating turns to take actions and a player has knowledge about what actions the other players have taken previously .
simultaneous games are usually represented in normal form , and sequential games are usually represented in extensive form .
a sequential game is considered a game of _ perfect information _ if all players know all the actions previously taken by the other players .
a similar concept is a game of _ complete information _ , which means all players in the game know the strategies and payoff functions of the other players . in some sense
, complete information may be viewed as capturing static information about a game while perfect information addresses dynamic information that becomes available during runs ( or instances ) of the game .
there are various solutions to a normal - form game depending upon the properties of the game and preferences over outcomes . a strategy , @xmath24 ,
is said to be _ dominant _ if it always results in higher payoffs than any other choice no matter what the opponents do , i.e. , @xmath45 in the example of the prisoner s dilemma game , the choice * defect * dominates * cooperate * for either player , though ironically both will be better off if they choose to cooperate and know that the other will also . in many games there are , however , no dominant strategies . to conservatively guarantee the best worst - case outcomes , a player may play the _ minimax _ strategy , which is @xmath46 where @xmath47 and @xmath48 are respectively a joint action for all players except @xmath8 and the set of joint actions for them . in theory , this can be solved via linear programming , but clearly there are many games that are too large to be solved in practice .
another approach to solving the problem is to find the _
best response _ strategies to the strategies of the other players . these can be defined as @xmath49 a joint strategy forms a _ nash equilibrium _ or if each individual strategy is the best response to the others strategies .
when a is reached , no player can be better off unilaterally , given that the other players stay with their strategies . in the example of the prisoner s dilemma game , * @xmath50defect , defect@xmath51 * is a .
although nash @xcite showed that all finite normal - form games have at least one , nash equilibria are generally difficult to achieve . on the one hand , conitzer and sandholm @xcite proved that computing nash equilibria is likely -hard ; on the other hand , some games involve more than one , thus without some extra coordination mechanism , no player knows which equilibrium the others would choose .
many papers have been concerned with equilibrium refinements " so as to make one equilibrium more plausible than another , however it seems to lead to overly complicated models that are difficult to solve .
a more practical approach is to allow players to learn by playing a game repeatedly .
a _ repeated game _ is a game made up from iterations of a single normal - form game , in which a player s strategy depends upon not only the one - time payoffs of different actions but also the history of actions taken by its opponents in preceding rounds .
such a game can be viewed as a system with multiple players and a single state , since the game setting does not change across iterations .
if the setting changes over time , the game becomes a _
stochastic game_. a stochastic game involves multiple states and the player payoff functions relate to both their actions in each interaction and the current state .
the goal of a player in such a game is to maximize its long - term return , which is sometimes defined as the average of all one - time payoffs or the discounted sum of those payoffs .
brown @xcite introduced a learning method , called _ fictitious play _ , for games in which all the other players use stationary strategies . with this method
, the player in question keeps a record of how many times the other players have taken each action and uses the frequencies of actions to estimate the probabilities of actions in an opponent s strategy . then the player chooses a best - response strategy based on its belief .
if the player s belief converges , what it converges to and its own best - response strategy form a @xcite .
the method becomes flawed if players adopt non - stationary strategies and in some games the belief simply does not converge @xcite .
another promising approach is to analyze the situation with evolutionary methods .
this approach assumes there is a large population of individuals and each strategy is played by a certain fraction of these individuals .
then , given the distribution of strategies , individuals with better average payoffs will be more successful than others , so that their proportion in the population will increase over time .
this , in turn , may affect which strategies are better than others .
in many cases , the dynamic process will move to an equilibrium .
the final result , which of possibly many equilibria the system achieves , will depend on the initial distribution .
the evolutionary , population - dynamic view of games is useful because it does not require the assumption that all players are sophisticated and think the others are also rational , an assumption that is often unrealistic . instead , the notion of _ rationality _ is replaced with the much weaker concept of _ reproductive success_. a related concept , considering the overall outcome rather than individual payoffs , is _
pareto optimality_. a strategy profile , @xmath52 , is _ pareto optimal _ , or _
pareto efficient _ , if there exists no other strategy profile producing higher payoffs for all players , i.e. , @xmath53 in the prisoner s dilemma game , all pure strategy profiles except for * @xmath50defect , defect@xmath51 * are pareto optimal .
a pareto optimal outcome is highly desirable , but usually difficult to achieve .
self - interested players tend to take locally optimal actions that may not collectively be pareto optimal . in the prisoner s dilemma game , * @xmath50cooperate , cooperate@xmath51 * instead of the * @xmath50defect , defect@xmath51 * obviously causes both players to be better off . in games of incomplete information , to utilize the concept of , each player needs to maintain an estimate of the others strategies so as to come up with a best - response strategy , where bayes theorem is used to update or revise beliefs following interactions with opponents .
the concept of equilibrium therefore becomes _ bayesian nash equilibrium _ , or . that is each player
s strategy is a function of her own information , and maximizes her expected payoff given other players strategies and given her beliefs about other players information @xcite .
auctions are a way to enable interactions among traders , and traders make profits as a result of transactions .
vickrey @xcite pioneered the approach of thinking about a market institution as a game of incomplete information since traders do not know each others private values . in research on single - sided auctions ,
the main goal is to find mechanisms that maximize the profit of sellers , who are special players in the auctioning games .
while in double - sided auctions , research focuses on maximizing social welfare and identifying how price formation develops dynamically . by assuming a fixed number of symmetric " , risk - neutral bidders , who each want a single unit of goods , have a private value for the object , and bid independently , vickrey showed that the seller can expect equal profits on average from all the standard types of auctions
this finding is called the _ revenue equivalence theorem_. this theorem provides the foundation for the analysis of _ _ optimal auctions _ _ and much subsequent research can be understood in terms of this theorem .
numerous articles have reported how its results are affected by relaxing the assumptions behind it .
the assumption that each trader knows the value of the goods being traded , and that these values are all private and independent of each other is commonly called the _ private - value model _ @xcite . in some cases , by contrast , the actual value of the goods is the same for everyone , but bidders have different private information about what that value actually is . in these cases ,
a bidder will change her estimate of the value if she learns another bidder s estimate , in contrast to the private - value case in which her value would be unaffected by learning any other bidder s preferences or information .
this is called the _ pure common - value model _ @xcite .
the winner in this scenario is the individual who makes the highest estimate of the value , and this tends to be an overestimate of the value .
this overestimation is called the _ winner s curse_. if all the bidders have the existence of the winner s curse in mind , the highest bid in first - price auctions tends to be lower than in those second - price auctions , though it still holds that the four standard auctions are revenue - equivalent .
a general model encompassing both the private - value model and the pure common - value model as special cases is the _ correlated - value model _ @xcite .
this assumes that each bidder receives a private information signal , but allows each bidder s value to be a general function of _ all _ the signals .
receives signal @xmath54 and would have value @xmath55 if all bidders signals were available to her . in the private - value model @xmath55
is a function only of @xmath54 . in the pure common - value model @xmath56 for all @xmath8 and @xmath57 . ]
milgrom and weber analyzed auctions in which bidders have _ affiliated information _ , and showed that the most profitable standard auction is then the ascending auction .
myerson @xcite demonstrated how to derive optimal auctions when the assumption of symmetry fails .
maskin and riley @xcite considered the case of risk - averse bidders , in which case the first - price sealed - bid auction is the most profitable of the standard auctions . for practical reasons ,
it is more important to remove the assumptions that the number of bidders is unaffected by the auction design , and that the bidders necessarily bid independently of each other . according to @xcite , sealed - bid designs frequently ( but not always ) both attract a large number of serious bidders and are better at discouraging collusion than english auctions .
in contrast with simple single - sided auctions , where the goals of auction mechanism designers reflect the interests of the single seller , double - sided auctions aim to maximize the collective interests of all traders , or in other words , the social welfare , i.e. , the total surplus all traders earn in an auction .
numerous publications have reported theoretical assertions or empirical observations of high efficiency in a variety of double - sided auctions , and have discussed what leads to the maximization of social welfare .
chatterjee and samuelson @xcite made the first attempt to analyze double auctions considering a special case of the involving a single buyer and a single seller . in this auction
, the transaction price is set at the midpoint of the interval of market - clearing prices when the interval is non - empty .
they found linear bidding strategies which miss potential transactions with probability of 1/6 .
satterthwaite and williams @xcite analyzed a generalized version of this auction so - called _
@xmath58-double auction _ or @xmath58-which involves @xmath59 sellers and @xmath59 buyers , and sets the transaction price at other points in the interval of market - clearing prices .
they showed that in s the differences between buyers bids and true values are @xmath60 and foregone gains from trade are @xmath61 , so _ _ ex post _ _ inefficiency vanishes reasonably fast as the market gets larger .
wilson @xcite first studied the generalization of games of incomplete information to s , in particular , s in which each agent can trade at most one indivisible unit and , given the bids and asks , the maximum number of feasible trades are made at a price a fraction @xmath58 of the distance between the lowest and highest feasible market clearing prices .
he proposed a strategy for buyers and sellers in which a trader waits for a while before making bids or asks .
then the trader conducts a dutch auction until an offer from the other side is acceptable .
this strategy produces a nearly _ ex post _ efficient final allocation .
wurman _ et al .
_ @xcite carried out an incentive compatibility analysis on a which is assumed to have @xmath62 bids and @xmath63 asks .
they showed that the ( @xmath64)st - price ( or @xmath63th - lowest - price ) clearing policy is incentive compatible for single - unit buyers under the private - value model , as is the @xmath62th - price ( or ( @xmath65)st - lowest - price ) auction for sellers . the only way to get incentive compatibility for both buyers and sellers is for some party to subsidize the auction .
myerson and satterthwaite @xcite showed that there does not exist any bargaining mechanism that is individually rational , efficient , and bayesian incentive compatible for both buyers and sellers without any outside subsidies .
as friedman @xcite pointed out , though theoretically it is natural to model s as a game of incomplete information , the assumption of prior common knowledge in the incomplete information approach may not hold in continuous auctions or may involve incredible computational complexity .
this is because at every moment , a trader needs to compute expected utility - maximizing shouts based on the shout and transaction history of the auction and the length of time the auction has to go . on the other hand
, laboratory results have shown that outcomes are quite insensitive to the number of traders beyond a minimal two or three active buyers and two or three active sellers .
foregone gains by satterthwaite and williams @xcite may suggest that the coefficient of @xmath66 in the actual foregone gain function is small .
] moreover , parameter choices , which according to an incomplete information analysis , should greatly reduce efficiency in s had no such effect in recent laboratory tests @xcite .
due to the difficulty of applying game - theoretic methods to complex auction mechanisms , researchers from economics and computer science have turned to running laboratory experiments , studying the dynamics of price formation and how the surprisingly high efficiency is obtained in a where information is scattered between the traders .
the data on which researchers base their studies may come from three sources : ( 1 ) field data from large - scale on - going markets , ( 2 ) laboratory data from small - scale auctions with human subjects , and ( 3 ) computer simulation experiments .
field data has the most relevance to the real - world economy , but does not reveal many important values , e.g. , the private values of traders , and hence puts limits on what can be done .
the human subjects in laboratory experiments presumably inherit the same level of intelligence and incentive to make profit as in real markets , and the experiments are run under the rules that researchers aim to study . such experiments , however , are expensive in terms of time and money needed .
computer - aided simulation is a less expensive alternative and can be repeated as many times as needed .
however traders strategies are not endogenously chosen as in auctions with human traders , but are specified exogenously by the experiment designers , which raises the question of whether the conclusions of this approach are trustworthy and applicable to practical situations .
gode and sunder @xcite invented a zero - intelligence strategy ( or ) that always randomly picks a profitable price to bid or ask .
surprisingly , their experiments with s exhibit high efficiency despite the lack of intelligence of the traders .
thereafter , much more work followed this path , and has gained tremendous momentum , especially given that real - world stock exchanges are becoming automated , e - business becomes an everyday activity , and the internet reaches every corner of the globe .
smith pioneered the research falling into the so - called _ experimental economics _ field by running a series of experiments with human subjects @xcite .
the experimental results revealed many of the properties of s , which have been the basis and benchmark for much subsequent work .
smith showed that in many different cases even a handful of traders can lead to high allocative efficiency , and transaction prices can quickly converge to the theoretical equilibrium .
smith s experiments are set up as follows : * every trader , either a buyer or a seller , is given a private value .
the set of private values form the supply and demand curves . *
each experiment was run over a sequence of trading days , or periods , the length of which depend on how many traders are involved but are typically several minutes in duration .
different experiments may have different numbers of periods .
* for simplicity , in most experiments , a trader is allowed to make a transaction for the exchange of only a single commodity in each day . *
traders are free at any time to make a bid / ask or to accept a bid / ask .
* once a transaction occurs , the transaction price , as well as the two traders private values , are recorded . * for each new day , a trader may make up to one transaction with the same private value as before no matter whether she has made one in the previous day . thus the supply and demand curves each correspond to a trading day .
the experimental conditions of supply and demand are held constant over several successive trading days in order to give any equilibrating mechanisms an opportunity to establish an equilibrium over time , unless it is the aim to study the effect of changing conditions on market behavior . @xcite reports 10 experiments that we discuss below .
each experiment was summarized by a diagram showing the series of transactions in the order in which they occurred .
figure [ fig : smith - test1 ] gives one of smith s diagrams .
are straight line segments because it is assumed there that a large number of traders participate in the auction and thus the step - changes can be treated as infinitesimal . ] . in the right - hand part of the diagram , each tick represents a transaction , rather than a unit of physical time .
trading prices in most experiments have a striking tendency to converge on the theoretical prices , marked with a dashed line in figure [ fig : smith - test1 ] .
to measure the tendency to converge , smith introduced the coefficient of convergence , @xmath14 , from ( [ equ : alpha ] ) .
figure [ fig : smith - test1 ] shows @xmath14 tends to decline from one trading day to the next .
the equilibrium price and quantity of experiments 2 and 3 are approximately the same , but the latter , with the steeper inclination of supply and demand curves , converges more slowly .
this complies with the walrasian hypothesis that the rate of increase in exchange price is an increasing function of the excess demand at that price .
experiment 4 presents an extreme case with a flat supply curve , whose result also confirms the walrasian hypothesis , but it converges to a fairly stable price above the predicted equilibrium . in this experiment ,
a decrease in demand is ineffective in shocking the market down to the equilibrium .
the result shows that the equilibrium may depend not only on the intersection of the supply and demand schedules , but also upon the shapes of the schedules . a hypothesis aiming to explain the phenomenon is that the actual market equilibrium will be above the equilibrium by an amount which depends upon how large the _ _ buyers rent _ _ , price axis , and the demand curve .
] is relative to the _
sellers rent_. , price axis , and the supply curve . ]
experiment 7 , which was designed with the purpose of supporting or contradicting this hypothesis , shows slow convergence , complying with the walrasian hypothesis , but still exhibits a gradual approach to equilibrium .
it is concluded that a still smaller buyers rent may be required to provide any clear downward bias in the static equilibrium .
what s more , it seems quite unmistakable " that the bigger the difference between the buyers rent and sellers rent , the slower the convergence .
smith speculated that the lack of monetary payoffs to the experimental traders may have an effect on the markets . a strong measure to further test the hypothesis is to mimic real markets as exactly as possible by paying each trader a small return just for making a contract in any period , which according to some experiments induces faster convergence .
experiment 5 was designed to study the effect on market behavior of changes in the conditions of demand and supply . at some point in the experiment ,
new buyers were introduced resulting in an increase in demand . the eagerness to buy causes the trading price to increase substantially once the market resumes and the price surpasses the previous equilibrium .
experiment 6 was designed to determine whether market equilibrium was affected by a marked imbalance between the number of intra - marginal sellers and the number of intra - marginal buyers near the predicted equilibrium price .
the result confirmed the effect of a divergence between buyer and seller rent on the approach to equilibrium , but the lack of marginal sellers near the theoretical equilibrium did not prevent the equilibrium from being attained .
the change of decrease in demand at the end of the fourth trading day showed that the market responded promptly by showing apparent convergence to the new , lower , equilibrium .
in contrast to the previous experiments , the market in experiment 8 was designed to simulate an ordinary retail market , in which only sellers are allowed to enunciate offers , and buyers could only either accept or reject the offers of sellers .
due to the desire of sellers to sell at higher prices , the trading prices in the first period remained above the predicted equilibrium . but starting at the second period , the trading price decreased significantly and remained below the equilibrium , not only because the early buyers again refrained from accepting any high price offers , but also because the competition among sellers became more intense . later in the experiment ,
when the previous market pricing organization was resumed , exchange prices immediately moved toward equilibrium .
experiments 9 and 10 are similar to experiment 7 except that each trader is allowed to make up to 2 transactions with the assigned private value within each day .
the results showed that the increase in volume helps to speed up the convergence to equilibrium .
the same results were obtained even when demand was increased during experiment 9 .
smith s focus in @xcite was mainly on the convergence of transaction prices in different scenarios rather than directly examining why high efficiency is obtained .
however , high efficiency is usually the goal of a market designer . in a computerized world
, a question that arises naturally is whether smith s results can be replicated in electronic auctions . in smith
s experiments , as is traditional in real markets , the traders are human beings , but computer programs are supposed to be automatic and work without human involvement .
obviously humans are intelligent creatures , but programs are not , at least for the foreseeable future .
is it intelligence that contributes to the high efficiency of double auction markets , or is it something else ?
gode and sunder @xcite were among the first to address this question , claiming that no intelligence is necessary for the goal of achieving high efficiency ; so the outcome is due to the auction mechanism itself .
they reached this position having introduced two trading strategies : _ zero intelligence without constraint _ or zi - u and _ zero intelligence with constraint _ or zi - c .
zi - u , the more nave version , shouts an offer at a random price without considering whether it is losing money or not , while zi - c , which lacks the motivation of maximizing profit and picks a price in a similar way to zi - u , simply makes shouts that guarantee no loss .
it was shown that performs poorly in terms of making a profit , but generates high efficiency solutions , comparable to the human markets ( see table [ table : zi - eff ] ) and can be considered to place a lower bound on the efficiency of markets @xcite .
.mean efficiency of markets in gode and sunder s experiments .
originally as table 2 in @xcite . [ cols="<,^,^,^,^,^",options="header " , ] gode and sunder s experiments were setup with similar rules as in smith s .
they designed five different supply and demand schedules and tested each of them respectively with the three kinds of homogeneous traders , , , and human traders .
figure [ fig : zi - test4 ] presents what happened in one of their experiments .
prices in the market exhibit little systematic pattern and no tendency to converge toward any specific level , but on the contrary , prices in the human market , after some initial adjustments , settle in the proximity of the equilibrium price ( indicated by a solid horizontal line in all panels in figure [ fig : zi - test4 ] ) .
gode and sunder then raised the question : how much of the difference between the market outcomes with traders and those with human traders is attributable to intelligence and profit motivation , and how much is attributable to market discipline ?
they argue that , after examining the performance of the markets , it is market discipline that plays a major role in achieving high efficiency .
though in the market , the price series shows no signs of improving from day to day , and the volatility of the price series is greater than the volatility of the price series from the human market , the series converges slowly toward equilibrium within each day .
gode and suner s explanation is that it is due to the progressive narrowing of the opportunity sets of traders , e.g. , the set of intra - marginal traders . despite the randomness of ,
buyers with higher private values tend to generate higher offered prices and they are likely to trade with sellers earlier than those buyers further down the demand curve .
a similar statement also holds for sellers .
thus as the auction goes on , the upper end of the demand curve shifts down and the lower end of the supply curve moves up , which means the feasible range of transaction prices narrows as more commodities are traded , and transaction prices will converge to the equilibrium price . the fact that traders lack profit motivation and have only the minimal intelligence ( just enough to avoid losing money ) suggests that the market mechanism is the key to obtaining high efficiency .
gode and sunder s results were , however , questioned by cliff and bruten @xcite .
the latter agreed on the point that the market mechanism plays a major role in achieving high efficiency , but disputed whether in markets transaction prices will always converge on equilibrium price .
they argued that the mean or expected value of the transaction price distribution was shown quantitatively to get close to the equilibrium price only in situations where the magnitude of the gradient of linear supply and demand curves is roughly equal , and used this to infer that zero - intelligence traders are not sufficient to account for convergence to equilibrium .
cliff and bruten further designed an adaptive trading strategy called _ zero intelligence plus _ or zip .
like , traders make stochastic bids , but can adjust their prices based on the auction history , i.e. , rasing or lowering their profit margins dynamically according to the actions of other traders in the market . more specifically ,
traders raise the profit margin when a less competitive offer from the competition is accepted , and lower the profit margin when a more competitive offer from the competition is rejected , or an accepted offer from the other side of the market would have been rejected by the subject . at every step ,
the profit margin is updated according to a learning algorithm called the _ widrow - hoff delta rule _ in which a value being learned is adapted gradually towards a moving target , and the past targets leave discounting momentum to some extent .
cliff and bruten concluded that the performance of zip traders in the experimental markets is significantly closer to that of human traders than is the performance of traders , based on the observation that traders rapidly adapt to give profit dispersion ] where @xmath23 and @xmath24 are the actual and theoretical equilibrium profits of trader @xmath8 , @xmath67 . ]
levels that are in some cases approximately a factor of ten less than those of traders .
preist and van tol introduced a revised version of , which we call , and reported faster convergence to equilibrium and robustness to changes in parameter configuration @xcite .
other learning methods have been adopted to design even more complex trading strategies than and its variants .
roth and erev @xcite proposed a reinforcement - based stimuli - response strategy , which we call re .
traders adapt their trading behavior in successive auction rounds by using their profits in the last round as a reward signal .
gjerstad and dickhaut @xcite suggested a best - response - based strategy , which is commonly referred to as gd .
traders keep a sliding window of the history of the shouts and transactions and calculate the probabilities of their offers being accepted at different prices .
the traders use a cubic interpolation on the shouts and transaction prices in the sliding window in order to compute the probability of future shouts being accepted .
they then use this to calculate the expected profit of those shouts .
the expected profit at a price is the product of the probability of the price being accepted and the difference between the price and the private value .
traders then always choose to bid or ask at a price that maximizes their expected profit .
is the most computation - intensive trading strategy considered so far , and indeed generates the best record both for allocative efficiency and the speed of convergence to equilibrium compared to the other trading strategies in literature
. by way of indicating typical efficiencies achieved in a cda , figure [ fig : eff ] shows the trend of the overall efficiencies of homogeneous cdas lasting 10 days with 50 rounds per day in which 10 buyers and 10 sellers all use the same strategy , one of : tt , , for more information .
] zip , re , and gd . the results are averaged over 400 iterations and obtained in jasa the extensible java - based auction simulation environment @xcite .
figure [ fig : ds ] gives the supply and demand schedules in the markets .
all the above empirical works have employed either human traders or homogeneous trading agents , demonstrating high efficiency and fast convergence to equilibrium , and some of this work has also produced theoretical results .
it is however necessary to see how an auction works populated by heterogeneous trading subjects .
there are both theoretical and practical reasons for considering heterogeneous traders . as rust _
argued in @xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ although current theories of markets have provided important insight into the nature of trading strategies and price formation , it is fair to say that none of them has provided a satisfactory resolution of hayek s problem " . in particular , current theories assume a substantial degree of implicit coordination by requiring that traders have common knowledge of each other s strategies ( in game - theoretic models ) , or by assuming that all traders use the same strategy ( in learning models ) .
little is known theoretically about price formation in markets populated by heterogeneous traders with limited knowledge of their opponents . ...
the assumption that players have common knowledge of each other s beliefs and strategies ... presumes an unreasonably high degree of implicit coordination amongst the traders ... game theory also assumes that there is no _ a priori _ bound on traders ability to compute their strategies .
however , even traders with infinite , costless computing capabilities may still decide to deviate from their strategies if they believe that limitations of other traders force them to use a sub - optimal strategy .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ they went on to argue that and other strategies striking performance strongly suggests that the nice properties have more to do with the market mechanism itself than the rationality of traders .
in addition , strategies that are more individually rational than may display less collective rationality since clever strategies can exploit unsophisticated ones such as and so that a more - intelligent extra - marginal trader has more chances to finagle a transaction with an intra - marginal traders , causing market efficiency to fall . to observe heterogeneous auctions , the santa fe double auction tournament ( ) was held in 1990 and prizes were offered to entrants in proportion to the trading profits earned by their programs over the course of the tournament .
30 programs from researchers in various fields and industry participated .
the majority of the programs encoded the entrant s market intuition " using simple rules of thumb .
the top - ranked program was , named after the entrant . and the runner - up strategy are remarkably similar . both
wait in the background and let the others do the negotiating , but when bid and ask get sufficiently close , jump in and steal the deal " @xcite .
the overall efficiency levels in the markets used in the tournaments originally appear to be somewhat lower than that observed in experimental markets with human traders , but experiments without the last - placed players produced an efficiency of around 97% .
this is further evidence that the properties of traders also affect the outcome of markets to some extent .
besides high efficiency levels and convergence to competitive equilibrium , other stylized facts " of human markets observed in the include : reductions in transaction - price volatility and efficiency losses in successive trading days that seem to reflect apparent learning effects , coexistence of extra - marginal and intra - marginal efficiency losses , and low - rank correlations between the _ realized order of transactions _ and the _ efficient order_. thorough examination of efficiency losses in the tournaments and later experiments indicates that the success of is due to its patience in waiting to exploit the intelligence or stupidity of other trading strategies .
the volume of e - commerce nowadays creates another motivation for evaluating trading strategies in a heterogeneous environment .
electronic agents , on behalf of their human owners , can automatically make strategic decisions and respond quickly to the changes in various kinds of markets . in the foreseeable future , these agents will have to compete with a variety of agents using a range of trading strategies and human traders . as more complex trading strategies appear , it is natural to speculate on how these electronic minds will compete against their human counterparts .
_ ran a series of s allowing persistent orders populated by a mixed population of automated agents ( using modified and strategies ) and human traders @xcite .
they found that though the efficiency of the s was comparable with prior research , the agents outperformed the humans in all the experiments , obtaining about 20% more profit .
_ speculated that this was due to human errors or weakness , and human traders were observed to improve their performance as they got familiar with using the trading software .
et al . _ also suggested that the weaknesses of trading agents may be found when human experts take them on and thus improvement can be made to the algorithms of the trading agents .
tesauro and das @xcite executed experiments with both homogeneous and heterogeneous trading agents with varying trader population composition , making it possible to gain more insights into the relative competitiveness of trading strategies . in either the so - called one - in - many " tests or balanced - group " tests , and ( and their variants ) exhibited superior performance over and even when the market mechanisms vary to some extent .
furthermore , , a variant of due to das _
@xcite , outperformed all the other strategies .
the above approaches nevertheless all employ a fixed competition environment . in practice ,
when a strategy dominates others , it tends to flourish and be adopted by more people .
are the first that we are aware of to conduct evolutionary experiments , where the relative numbers of the different trading strategies changed over time , so that more profitable strategies became more numerous than less profitable ones .
such an analysis revealed that although agents outperformed others when traders of different types are approximately evenly distributed , they later exhibited low overall efficiency as they became the majority , making the evolution process a cycle of ups and downs .
walsh _ et al . _
@xcite gave a more formal analysis combining the game - theoretic solution concept of and _ replicator dynamics_. they treated heuristic strategies , rather than the atomic actions like a bid or ask , as primitive , and computed expected payoffs of each individual strategy at certain points of the joint heuristic strategy space .
this method reduced the model of the game from a potentially very complex , multi - stage game to a one - shot game in normal form . at points where one strategy gains more than others , replicator dynamics dictates that the whole population moves to a nearby point where the winning strategy takes a larger fraction of the population .
this process continues until an equilibrium point is reached where either the population becomes homogeneous or all strategies are equally competitive in terms of their expected payoffs .
there may be multiple equilibrium points ` absorbing ' areas of different sizes , _ basins _ of the equilibria , which together compose the whole strategy space .
in particular , figure [ fig : walsh - before - perturbation ] shows the replicator dynamics of a market with three strategies .
@xmath68 , @xmath69 , @xmath70 , and @xmath71 are all equilibrium points , but @xmath69 and @xmath71 are not stable since a small deviation from them will lead to one of the other equilibria .
the triangle field gives an overview of the interaction of the three strategies and their relative competitiveness .
what s more , a technique called _ perturbation analysis _ is used to evaluate the potential to improve on a strategy .
figure [ fig : walsh - after - perturbation ] shows the replicator dynamics of the same strategies after small portions of both and s payoffs were shifted to .
such a shift significantly changed the landscape of the space , and dominated in most of possible combinations .
this showed that a ` tiny ' improvement on the strategy may greatly affect its competition against the other strategies .
phelps _ et al . _
@xcite took a similar approach in comparing the , , and strategies , showed the potential of , and demonstrated that a modified strategy could be evolved by optimizing its learning component .
the main drawback of this approach is an exponential dependence on the number of strategies , which limits its applicability to real - world domains where there are potentially many heuristic strategies .
@xcite proposed information theoretic approaches to deliberately choose the sample points in the strategy space through an interleaving of equilibrium calculations and payoff refinement , thus reducing the number of samples required .
designing heuristic strategies to a great extent depends on the intelligence and experience of the strategy designer .
prior studies have also demonstrated that heuristic strategies performance hinges on the selection of parameter values .
automatic optimization is preferable in this sense to find best parameter combinations and further identify better strategies .
cliff and phelps _ et al .
_ are the pioneers in this work , adopting evolutionary computation to address the challenge .
a _ genetic algorithm _ ( or ) is a search technique used in computing to find true or approximate solutions to optimization and search problems .
genetic algorithms are a particular class of evolutionary algorithms that use techniques inspired by evolutionary biology such as _ inheritance _ , _ mutation _ , _ selection _ , and _ crossover _ @xcite .
a typical genetic algorithm requires two things to be defined : 1 .
a _ genetic representation _ of the solution domain , also called the _ genotype _ or _ chromosome _ of the solution species , 2 . a _ fitness function _ to evaluate the solution domain .
a standard representation of the solution is as an array of bits .
arrays of other types and structures can be used in essentially the same way . the main property that makes these genetic representations convenient
is that their parts are easily aligned due to their fixed size , which facilitates simple crossover operation .
variable length representations have also been used , but crossover implementation is more complex in this case .
the fitness function is defined over the genetic representation of a solution and measures the quality of the solution .
the fitness function is always problem dependent .
for instance , in the knapsack problem we want to maximize the total value of objects that we can put in a knapsack of some fixed capacity .
a representation of a solution might be an array of bits , where each bit represents a different object , and the value of the bit ( 0 or 1 ) represents whether or not the object is in the knapsack .
not every such representation is valid , as the size of objects may exceed the capacity of the knapsack .
the fitness of the solution is the sum of values of all objects in the knapsack if the representation is valid , or 0 otherwise . in some problems ,
it is hard or even impossible to define the fitness expression ; in these cases , interactive genetic algorithms are used .
once we have the genetic representation and the fitness function defined , the proceeds to initialize a population of solutions randomly , then improve it through repetitive application of mutation , crossover , and selection operators .
cliff addressed the labor - intensive manual parameter optimization for the strategy , automatically optimizing parameter selection using a @xcite .
he identified eight parameters in : lower and upper bounds of the learning rate @xmath72 ( how fast to move towards the target ) , momentum @xmath73 ( how much past momentum to carry over ) , and initial profit margin @xmath74 , and the upper bounds of the ranges defining the distributions of absolute and relative perturbations on learned prices , respectively denoted as @xmath75 and @xmath76 .
these real parameters make an eight - dimensional space and any parameter value combination corresponds to a point in that space .
the vector of the eight parameters defines an ideal genotype .
phelps _ et al .
_ took a step further along this track .
they combined the replicator - dynamics - based heuristic strategy analysis method in @xcite and a , identified a strategy as the basis for optimization , and successfully evolved the strategy and acquired an optimized strategy that can beat , commonly considered the most competitive strategy @xcite . since it is not realistic to seek best " , or even good " , strategies that can beat all potential opponents because an absolutely dominating strategy does not appear to exist in the trading scenario
since the performance of a strategy depends greatly on the types of the opponents
proposed using a small finite population of randomly sampled strategies to approximate the game with an infinite strategy population consisting of a mixture of all possible strategies . in particular , , , and were chosen as sample strategies . following the heuristic strategy analysis and perturbation method in @xcite , was found to have the potential to dominate and .
the strategy uses reinforcement learning to choose from @xmath17 possible profit margins over the agent s private value based on a reward signal computed as a function of profits earned in the previous round of bidding .
potentially , the learning algorithm may be replaced by a number of learning algorithms , including ( stateless q - learning ) , ( a modified version of used in @xcite ) , and ( a control algorithm which selects a uniformly random action regardless of reward signal ) .
then encoded the genotype to select any of these algorithms together with their parameters .
the evolutionary search procedure they used is similar to cliff s except that the individuals in a generation are evaluated again with the heuristic strategy analysis approach and the basin size is used as a measure of fitness .
the experiment finally found a algorithm with a particular parameter combination , which together with composes the nash equilibrium that captures 97% of the strategy space populated by the learned strategy , , , and . the trading agent competition ( )
was organized to promote and encourage high quality research into trading agents . under the umbrella ,
a series of competitions have been held , including two types of game , lassic and @xcite .
lassic sets up a travel agent " scenario based on complex procurement in multiple simultaneous auctions .
each travel agent ( an entrant to the competition ) has the goal of assembling travel packages ( from town to tampa , during a notional multi - day period ) .
each agent is acting on behalf of a certain number of clients , who express their preferences for various aspects of the trip .
the objective of the travel agent is to maximize the total satisfaction of its clients ( the sum of the client utilities ) .
was designed to capture many of the challenges involved in supporting dynamic supply chain practices in the industry of pc manufacturing .
supply chain management is concerned with planning and coordinating the activities of organizations across the supply chain , from raw material procurement to the delivery of finished goods . in today
s global economy , effective supply chain management is vital to the competitiveness of manufacturing enterprizes as it directly impacts their ability to meet changing market demands in a timely and cost effective manner .
in , agents are simulations of small manufacturers , who must compete with each other for both supplies and customers , and manage inventories and production facilities .
mechanism design applied to auctions explores how to design the rules that govern auctions to obtain specific goals .
the story of trading strategies in the preceding section is only one facet of the research on auctions .
gode and sunder s results suggest that auction mechanisms play an important role in determining the outcome of an auction , and this is further bourne out by the work of walsh _ et al . _
@xcite , which also points out that results hinge on both auction design and the mix of trading strategies used . according to classical auction theory ,
if an auction is _ strategy - proof _ or _ incentive compatible _ , traders need not bother to conceal their private values and in such auctions complex trading agents are not required .
however , typical s are not strategy - proof .
mcafee @xcite has derived a form of double auction that is strategy - proof , though this strategy - proofness comes at the cost of lower efficiency . despite the success of analytic approaches to the relatively simple auctions presented in section [ sec : auction - theory ] ,
the high complexity of the dynamics of some other auction types , especially s , makes it difficult to go further in using analytical methods @xcite . as a result
, researchers turned to empirical approaches using machine learning techniques , sometimes combined with methods from traditional game theory . instead of trying to design optimal auction mechanisms ,
the computational approach looks for relatively good auctions and aims to make them better , in a noisy economic environment with traders that are not perfectly rational .
one can think of different forms of auctions as employing variations of a common set of the auction rules , forming a parameterized auction space .
et al . _ and others parameterized auction rules using the following classification @xcite : * bidding rules : determine the semantic content of messages , the authority to place certain types of bids , and admissibility criteria for submission and withdrawal of bids . *
* how many sellers and buyers are there ? * * are both groups allowed to make shouts ? * * how is a shout expressed ? * * does a shout have to beat the corresponding market quote if one exists ? * information revelation : * * when and what market quotes are generated and announced ? * * are shouts visible to all traders ?
* clearing policy : * * when does clearing a market take place ? * * when does a market close ? * * how are shouts matched ? * * how is a transaction price determined ? the idea of parameterizing auction space not only eases the heuristic auction mechanism design , but also makes it possible to ` search ' for better mechanisms in an automated manner @xcite .
it is not yet clear how auction design , and thus the choice of parameter values , contributes to the observed performance of auctions .
thus it is not clear how to create an auction with a particular specification . it _ is _ possible to design simple mechanisms in a provably correct manner from a specification , as shown by conitzer and sandholm @xcite .
however it is not clear that this kind of approach can be extended to mechanisms as complex as s. as a result , it seems that we will have to design double auction mechanisms experimentally , at least for the foreseeable future .
of course , doing things experimentally does not solve the general problem .
a typical experimental approach is to fix all but one parameter , creating a one - dimensional space , and then measure performance across a number of discrete sample points in the space , obtaining a fitness landscape that is expected to show how the factor in question correlates to a certain type of performance and how the auction can be optimized by tweaking the value of that factor @xcite . in other words ,
the experimental approach examines one small part of a mechanism and tries to optimize that part .
the situation is complicated when more than one factor needs to be taken into consideration the search space then becomes complex and multiple dimensional , and the computation required to map and search it quickly becomes prohibitive . instead of manual search
, some researchers have used evolutionary computation to automate mechanism design in a way that is similar to the evolutionary approach to optimizing trading strategies .
cliff @xcite explored a continuous space of auction mechanisms by varying the probability of the next shout ( at any point in time ) being made by a seller , denoted by @xmath77 .
the continuum includes the cda ( @xmath78 ) and also two purely single - sided mechanisms that are similar to the english auction ( @xmath79 ) and the dutch auction ( @xmath80 ) .
cliff s experiments used genetic algorithms and found that a @xmath77 that corresponds to a completely new kind of auction led to a better @xmath14 value than that obtained for other markets using zip traders .
_ @xcite and the same authors but in a different order @xcite continued with this work , showing that the approach is also effective in markets using zi - c traders , and the new irregular " mechanisms can lead to high efficiency with a range of different supply and demand schedules as well . the visualization of fitness landscapes , using plots including 3d histograms and contours , is also noteworthy .
byde @xcite took a similar approach in studying the space of auction mechanisms between the first and second - price sealed - bid auctions .
the winner s payment is determined as a weighted average of the two highest bids , with the weighting determined by the auction parameter . for a given population of bidders ,
the revenue - maximizing parameter is approximated by considering a number of parameter choices over the allowed range , using a to learn the parameters of the bidders strategies for each choice , and observing the resulting average revenues . for different bidder populations ( varying bidder counts , risk sensitivity , and correlation of signals ) ,
different auction parameter values are found to maximize revenue . taking another tack , phelps _ et al .
_ explored the use of genetic programming to determine auction mechanism rules automatically .
_ genetic programming _ ( or ) , another form of evolutionary computation that is similar to s , evolves programs ( or expressions ) rather than the binary strings evolved in s. this makes automatic programming possible , and in theory allows even more flexibility and effectiveness in finding optimal solutions in the domain of concern . in
, programs are traditionally encoded as _ tree structures_. every tree node has an operator function and every terminal node has an operand , making mathematical expressions easy to evolve and evaluate . with tree structures ,
crossover is applied on an individual by simply switching one of its nodes with another node from another individual in the population .
mutation can replace a whole node in the selected individual , or it can replace just the information of that node . replacing
a node means replacing the whole branch .
this adds greater effectiveness to the crossover and mutation operators @xcite .
phelps _ et al .
_ @xcite demonstrated how can be used to find an optimal point in a space of pricing policies , where the notion of optimality is based on allocative efficiency and trader market power . in markets , there are two popular pricing policies : the @xmath58- pricing rule @xcite and the uniform pricing policy .
the former is clearly a discriminatory policy and may be represented as : @xmath81 where @xmath82 $ ] , and @xmath83 and @xmath84 are ask and bid prices . the latter executes all transactions at the same price , typically the middle point of the interval between the market ask and bid quotes . searching in the space of arithmetic combinations of shout prices and market quotes including the above two rules as special cases , led to a complex expression that is virtually indistinguishable from the @xmath85 version of the @xmath58- pricing rule .
this shows that the middle - point transaction pricing rule not only reflects the traditional practice but also can be technically justified . noting that the performance of an auction mechanism always depends on the mix of traders participating in the mechanism , and both the auction mechanism and the trading strategies may adapt themselves simultaneously , phelps _ et al . _ @xcite further investigated the use of co - evolution in optimizing auction mechanisms .
they first co - evolved buyer and seller strategies and then together with auction mechanisms .
the approach was able to produce outcomes with reasonable efficiency in both cases .
phelps _ et al .
_ proposed a novel way to evaluate and compare the performances of market mechanisms using heuristic strategy analysis @xcite . despite the fact that the performance of an auction mechanism may vary significantly when the mechanism engages different sets of trading agents , previous research on auctions analyzed the properties of markets using an arbitrary selection of homogeneous trading strategies .
a more sound approach is to find the equilibria of the game between the participating trading strategies and measure the auction mechanism at those equilibrium points . as sections [ sec :
trading : interaction - heterogeneous ] and [ sec : trading : automating : phelps ] have discussed , the heuristic strategy analysis calculates equilibria among a representative collection of strategies .
this makes the method ideal for measuring market mechanisms at those relatively stable equilibria .
the representative strategies selected by phelps _
included , , and .
the replicator dynamics analysis revealed that : ( 1 ) neither the nor the mechanism is strategy - proof since is not dominant in either market ; ( 2 ) increasing the number of agents in the led to the appearance of an equilibrium basin for an equilibrium near , which agreed with the conclusion drawn through the approximate analysis in @xcite discussed in section [ sec : auction - theory : da ] ; and ( 3 ) the has higher efficiency than the in the sense that the three equilibrium points in the dynamics field for the all generate 100% efficiency while the only equilibrium for produces 98% efficiency .
one can interpret the small efficiency difference as justifying the s use of a rather than a for faster transactions and higher volumes .
one avenue of future research is to combine this evaluation method with evolutionary computation to optimize mechanisms . considering that the information about the population of traders is usually unknown to the auction mechanism , and many analytic methods depend on specific assumptions about traders , pardoe and stone advocated a self - adapting auction mechanism that adjusts auction parameters in response to past auction results @xcite .
their framework includes an _ evaluator _ module , which can create an auction mechanism for online use , can monitor the performance of the mechanism , and can use the economic properties of the mechanism as feedback to guide the discovery of better parameter combinations .
this process then creates better auction mechanisms that continue to interact with traders which are themselves possibly evolving at the same time . a classic algorithm for @xmath17-armed bandit problems , @xmath86-greedy ,
is used in the evaluator module to make decisions on parameter value selection .
this work differs from previous work in the sense that here auction mechanisms are optimized during their operation while the mechanisms in the approaches discussed before find remain static and are assumed to perform well even when they face a set of traders that is different from those used in searching for the mechanisms . following the lassic and the competitions introduced in section [ sec : trading : tac ] , a new competition called
was run in the summer of 2007 in order to foster research on auction mechanism design . in ,
the software trading agents are created by the organizers of the competition , and entrants compete by defining rules for matching buyers and sellers and setting commission fees for providing this service .
entrants compete against each other in attracting buyers and sellers and making profits .
this is achieved by having effective matching rules and setting appropriate fees that are a good trade - off between making profit and attracting traders .
we developed @xcite , based on phelps s , to run as the game server .
it provides various trading strategies , market selection strategies , and market mechanism frameworks to avoid entrants working from scratch .
is also an ideal experimental platform for researchers to evaluate auction mechanisms in a competition setting .
this report aims to provide an overview of the field of auction mechanism design and build the foundation for further research .
auctions are markets with strict regulations where traders negotiate and make deals
. an auction may be single - sided or double - sided depending upon whether only sellers or only buyers can make offers or whether both can . the four standard single - sided auctions
english auction , dutch auction , first- and second - price sealed - bid auctions have been the subject of traditional auction theory .
vickrey s pioneering work in this area led to the revenue equilibrium theorem that shows a seller can expect equal profits on average from all the standard types of auctions with a few assumptions about the bidders .
other researchers followed the approach and managed to extend the applicability of the theorem when the assumptions are relaxed .
double - sided auctions , which are important in the business world , posed a bigger challenge due to the higher complexity of their structure and the interaction between traders . while classical mathematical approaches have continued to be successful in analyzing some simple types of double auctions , they have been unable to apply to more practical scenarios .
smith and others initiated experimental approaches and showed that double auctions , even with a handful of traders , may lead to high allocative efficiency and the transaction prices quickly converge to the expected equilibrium price .
subsequent experiments with human and/or artificial traders tried to explain what led to these desirable properties and tended to show that auction mechanisms played a major role , though the intelligence of traders had an effect as well
. further work , on the one hand , introduced more and more complex trading strategies not only making higher individual profits but also improving the collective properties of auctions . on the other hand ,
different methods have been explored to design novel auction mechanisms .
one approach is to evolve parameterized auction mechanisms based on evolutionary computation .
et al . _ have found a new variant of continuous double auctions through evolving mechanisms that converge more quickly to equilibrium , and also exhibit higher efficiency than those previously known .
et al . _ have explored the use of genetic programming and justified the traditional mid - point transaction pricing rule as optimizing efficiency while balancing trader market power .
in addition to these off - line techniques for optimization through evolutionary computing , online approaches have been proposed to produce adaptive auction mechanisms , which , with dynamic trader populations , can continuously monitor and improve their performance . with the understanding of this prior research work ,
what can be done further at the interface of computer science and economics include : obtaining more insights into double - sided auction mechanisms , inventing novel auction rules , and searching for optimal combinations of various kinds of policies , automatically producing desirable auction mechanisms .
dave cliff .
evolutionary optimization of parameter sets for adaptive software - agent traders in continuous double auction markets .
technical report , hewlett - packard research laboratories , bristol , england , 2001 .
vincent conitzer and tuomas sandholm .
automated mechanism design : complexity results stemming from the single - agent setting . in _ the 5th international conference on electronic commerce ( icec03 ) _ , pages 1724 , pittsburgh , pa , usa , september 2003 .
vincent conitzer and tuomas sandholm .
an algorithm for automatically designing deterministic mechanisms without payments . in _ third international joint conference on autonomous agents and multiagent systems - volume 1 ( aamas04 ) _ , pages 128135 , new york city , ny , usa , july 2004 .
rajarshi das , james e. hanson , jeffrey o. kephart , and gerald tesauro .
agent - human interactions in the continuous double auction . in
_ proceedings of the 17th international joint conference on artificial intelligence _ ,
seattle , u.s.a .
, august 2001 .
simon parsons , mark klein , and juan antonio rodriguez - aguilar .
a bluffer s guide to auctions .
technical report , center for coordination science , sloan school of management , massachusetts institute of technology , 2004 . research note .
steve phelps , marek marcinkiewicz , simon parsons , and peter mcburney .
using population - based search and evolutionary game theory to acquire better - response strategies for the double - auction market . in _ proceedings of ijcai-05 workshop on trading agent design and analysis ( tada-05 )
_ , 2005 .
steve phelps , marek marcinkiewicz , simon parsons , and peter mcburney .
a novel method for automatic strategy acquisition in n - player non - zero - sum games . in _ proceedings of the fifth international joint conference on autonomous agents and multi - agent systems ( aamas06 )
_ , pages 705712 , new york , ny , usa , 2006 .
acm press .
steve phelps , peter mcburney , simon parsons , and elizabeth sklar .
co - evolutionary auction mechanism design : a preliminary report . in
_ proceedings of workshop on agent mediated electronic commerce iv ( amec iv ) _
, 2002 .
steve phelps , simon parsons , and peter mcburney .
an evolutionary game - theoretic comparison of two double - auction market designs . in _ proceedings of workshop on
agent mediated electronic commerce vi ( amec vi ) _ , 2004 .
steve phelps , simon parsons , elizabeth sklar , and peter mcburney . using genetic programming to optimise pricing rules for a double auction market . in _ proceedings of the workshop on agents for electronic commerce , _ ,
pittsburgh , pa , 2003 .
chris preist and maarten van tol .
adaptive agents in a persistent shout double auction . in
_ proceedings of the 1st international conference on information and computation economies _ , pages 1118 .
acm press , 1998 .
william walsh , rajarshi das , gerald tesauro , and jeffrey o. kephart .
analyzing complex strategic interactions in multi - agent systems . in piotr gmytrasiewicz and simon
parsons , editors , _ proceedings of 2002 workshop on game - theoretic and decision - theoretic agents ( gtdt-02 ) _ , edmonton , alberta canada , july 2002 .
william e. walsh , david c. parkes , and rajarshi das .
choosing samples to compute heuristic - strategy nash equilibrium . in _
aamas 2003 workshop on agent mediated electronic commerce _ , melbourne , australia , 2003 . | auctions are markets with strict regulations governing the information available to traders in the market and the possible actions they can take . since well designed auctions achieve desirable economic outcomes , they have been widely used in solving real - world optimization problems , and in structuring stock or futures exchanges .
auctions also provide a very valuable testing - ground for economic theory , and they play an important role in computer - based control systems . auction mechanism design aims to manipulate the rules of an auction in order to achieve specific goals .
economists traditionally use mathematical methods , mainly game theory , to analyze auctions and design new auction forms .
however , due to the high complexity of auctions , the mathematical models are typically simplified to obtain results , and this makes it difficult to apply results derived from such models to market environments in the real world . as a result , researchers are turning to empirical approaches .
this report aims to survey the theoretical and empirical approaches to designing auction mechanisms and trading strategies with more weights on empirical ones , and build the foundation for further research in the field .
= 1 |
theoretical investigation of the cosmological inhomogeneity growth presents one of the most serious challenges to computational astrophysics .
a number of problems arises on non - linear scales when the density contrast can not be assumed small .
these include simulations of the galaxy formation process , working out predictions concerning the galactic environment ( number of dwarf satellite galaxies ) and structure of their central regions ( the cusp - core problem ) either with cold dark matter ( dm ) or within warm dm models @xcite .
interesting possibility to obtain bounds on masses of dm particles stems from observations of ly-@xmath3 forest ( see , e.g. , @xcite ) and references therein ) , which requires accurate calculations of the power spectrum of cosmological inhomogeneity on kiloparsec scales .
most developed computational techniques to study the cosmological structure formation involve n - body simulations combined with the smoothed particle hydrodynamics @xcite .
currently performed simulations involve up to @xmath4 particles ( see , e.g. , @xcite ) .
these methods should be tested in independent simulations . on the other hand ,
some analytical and semi - analytical schemes were proposed to study the matter power spectra on small scales @xcite .
these typically involve perturbative schemes after transition to the fourier - transformed hydrodynamical variables .
there are techniques dealing with correlation functions in the fourier space @xcite .
these authors use some additional _ _ a__priori suggestions in order to close the infinite chain of correlation functions .
the validity of these suggestions is not evident .
a comparison of different approaches can be found in @xcite . in this paper
we propose an alternative method , which uses a direct integration of hydrodynamical equations in the fourier space . in order to estimate workability of the method , as a first step we consider a one - dimensional problem of hydrodynamical evolution for a pressureless gravitating matter , i.e. one - dimensional density shells .
the integration is performed for each realization from a random ensemble of initial data with subsequent averaging procedure . in order to test the numerical simulations
, we found an analytical solution which describes one - dimensional collapse of the plane shells ( section [ analytical_solution ] ) .
this solution is used in case of periodic initial data corresponding to a symmetric motion . in section [ numerical_simulations ]
we write down the equations for the fourier coefficients and present some results for the power spectrum obtained after the statistical averaging .
in order to test numerical simulations , it is useful to have an exact solution . in this section
we obtain such a solution of one - dimensional problem in the lagrange variables .
the 1-dimensional version of hydrodynamical equations ( continuity , euler and poisson equations ) in case of a pressureless gravitating fluid is : @xmath5 @xmath6 @xmath7 @xmath8 is the velocity , @xmath9 is the mass density , @xmath10 is the gravitational potential , @xmath11 is the gravitational constant . in order to pass to the lagrangian description we introduce the stream lines @xmath12 : @xmath13 equation ( [ euler ] ) yields @xmath14
let for @xmath15 : @xmath16 . using ( [ continuity ] ) , ( [ euler ] ) we see that @xmath17 is constant along the stream lines .
this is a consequence of mass conservation in plane layer between @xmath18 and @xmath19 .
then we obtain because of ( [ eq8 ] ) @xmath20 - 2\pi g\,t^2\int\limits_{\xi _ 2 } ^{\xi _ 1 } { dx } \,\rho _ 0 ( x).\ ] ] formula ( [ eq9 ] ) allows to get a general solution of the one - dimensional problem in lagrangian coordinates .
further for simplicity we deal with a symmetric motion ( with respect to the coordinate origin @xmath21 ) so that @xmath22 , @xmath23 . in this case
@xmath24 the solution is valid during a limited time until @xmath25 transition to euler variables is @xmath26 where @xmath27 is defined implicitly as the solution of the equation : @xmath28 in case of homogeneous initial conditions @xmath29 , @xmath30 , @xmath31 is the `` hubble constant '' , we get : @xmath32 @xmath33 or @xmath34 to consider deviations from the homogeneous background , we impose periodic boundary conditions such that the initial conditions ( @xmath35 ) are as follows : @xmath36 , \,\ , \rho_{0 } ( \xi ) = \rho_{h0 } \left [ { 1 + \sum\limits_{\substack{n=1 } } ^\infty { b_{n}^{\,0 } } \cos ( k_{n } \xi ) } \right],\ ] ] where @xmath37 is the periodic `` box '' size , @xmath38 .
substitution of the initial conditions into ( [ eq10]-[eq12 ] ) yields @xmath39 @xmath40}{r(t)+th_0 \sum\limits_{n=1}^{\infty } a_n^0 \cos(k_n\xi)-\frac{3}{4}\left(t\,h_0\right)^2 \sum\limits_{n=1}^{\infty } b_n^{\,0 } \cos(k_n\xi)},\ ] ] the solution shows a singular growth ( collapse of one - dimensional gravitating layers ) as the condition ( [ eq11 ] ) is violated . at the end of this section
we note that in spite of the exact form of the solution given by ( [ eq10]-[eq12 ] ) we can not avoid a numerical work when we pass to the euler variables and then to the fourier representation .
so we refer to the method of this section as `` semi - analytical '' .
furthermore for the homogenous background we denote @xmath41 , @xmath42 is the `` conformal time '' : @xmath43 , and @xmath44 is the comoving spatial coordinate . after some calculation on account of ( [ eq14 ] )
we get @xmath45,\,\ , \mathcal h(\tau ) = \frac{dr}{dt}= \frac{1}{r}\frac{dr}{d\tau } = - 2h_0 \tanh ( h_0 \tau ) .\ ] ] hereafter @xmath46 is the density contrast , @xmath47 , @xmath48 is the peculiar velocity . taking into account poisson equation ( [ eq6 ] )
we get : @xmath49 in terms of conformal @xmath42 and comoving @xmath50 the hydrodynamic equations can be written as : @xmath51 @xmath52 we proceed to deal with the fourier coefficients in the symmetric one - dimensional case .
@xmath53 @xmath54 the reverse transformation is : @xmath55 we assume @xmath56 at @xmath57 then it is easy to see from ( [ eq16],[eq17 ] ) that these equalities are fulfilled for all @xmath58 .
the equations for the fourier coefficients take on the form : @xmath59 @xmath60 the numerical solution of the equations ( [ eq21 ] ) , ( [ eq22 ] ) was performed using the 4-th order runge - kutta method after a truncation of the infinite chain of coefficients @xmath61 .
calculations were carried out by the specially written gpgpu code using opencl sdk by amd .
the time , required to calculate @xmath62 and @xmath63 , in case of 256 values of @xmath64 ( points over @xmath65 ) for single realization of initial conditions , is 10 seconds .
this is considerably faster than direct using of the semi - analytical solution of section [ analytical_solution ] .
it is important to note that it is easy to extend the corresponding algorithms to the 3-d case .
we calculated coefficients @xmath63 of the density contrast as a function of @xmath66 by means of analytical and numerical methods with the same initial conditions . on the fig .
[ fig1 ] these coefficients are presented for @xmath67 and @xmath68 for both methods . for larger @xmath66
we observe an infinite growth ( for finite time ) that corresponds to collapse of some plane gravitating shells due violation of condition ( [ eq11 ] ) .
correspondingly , the difference between two methods , that reflects the error of calculation , increases for greater @xmath66 and greater @xmath64 .
for example , in order to look how the perturbation propagates from small wavenumbers to larger ones , we considered the following initial conditions : @xmath69 where integer @xmath64 varies @xmath70 to @xmath71 except @xmath72 ; @xmath73 ; all @xmath74 . for @xmath75 or @xmath76
the difference between the values calculated by different methods is less than 1% , and for @xmath68 the difference changes from 1.5% to 6.7% as @xmath38 increases from 0.6 to 6 .
larger @xmath65-interval is presented on fig .
the next figure presents the power spectrum obtained by averaging of the solutions for the ensemble of initial data with uniform distribution of @xmath77 , @xmath78 .
we observe the growth of dispersion , which is explained as follows : as @xmath66 grows , some of realizations of the ensemble of solutions ( with larger @xmath79 ) enter the region which is close to the singularity . for @xmath80 ( top ) and @xmath81 ( bottom )
determined by semi - analytical and numerical methods with the only nonzero initial values @xmath82.,title="fig:",width=529 ] -3 mm for the initial conditions as described on fig .
, title="fig:",width=529 ] -3 mm calculated for 100 realizations of randomly generated independent initial conditions .
, title="fig:",width=529 ] -3 mm
we present a new approach to investigation of the cosmological inhomogeneity by means of the direct integration of hydrodynamic equations in the fourier space . at the moment we studied a one - dimensional hydrodynamical evolution of cold ( pressureless ) gravitating matter .
the numerical integration has been fulfilled for a random ensemble of initial conditions with subsequent averaging procedure to get the power spectrum of the density contrast .
we used the gpgpu instead of the classic cpu because the problem can be easily processed in parallel environment .
the numerical simulations have been tested using the analytic solution that describes the one - dimensional collapse of gravitating shells .
the density contrast shows a propagation of perturbations from smaller wavenumbers to larger ones .
the evolution in time ends with a singular growth of the density contrast .
correspondingly , we point out a significant growth of dispersion of the power spectrum in the non - linear region .
we consider our results as a first step to the simulations of cosmological inhomogeneity growth in the cold matter that could be an alternative to the cosmological n - body simulations .
we expect that our approach will be especially effective in the weakly nonlinear regime
. the next step will be an implementation of the three - dimensional case of the problem , which is technically similar to the one dimensional problem .
the trial runs of our method allow us to think that it could be really used for power spectrum calculations in the 3-d case with realistic requirements to the computer time . | we propose a method of calculation of the power spectrum of cosmological perturbations by means of a direct numerical integration of hydrodynamic equations in the fourier space for a random ensemble of initial conditions with subsequent averaging procedure .
this method can be an alternative to the cosmological n - body simulations .
we test realizability of this method in case of one - dimensional motion of gravitating matter pressureless shells . in order to test the numerical simulations
, we found an analytical solution which describes one - dimensional collapse of plane shells .
the results are used to study a nonlinear interaction of different fourier modes . + _ key words : _ large scale structure , cosmological perturbations , hydrodynamics . @xmath0__astronomical observatory , national taras shevchenko university of kyiv , observatorna str.,3 , kiev 04053 , ukraine _ _
+ @xmath1 [email protected] + @xmath2 [email protected] |
nanotechnology has the potential to revolutionize health care @xcite . a current example is enhanced imaging with nanoscale particles @xcite .
future possibilities include programmable machines comparable in size to cells @xcite .
such microscopic robots ( `` nanorobots '' ) could provide significant medical benefits @xcite . realizing these benefits requires fabricating the robots cheaply and in large numbers .
such fabrication is beyond current technology , but could result from ongoing progress in developing nanoscale devices .
one approach is engineering biological systems , e.g. , rna - based logic inside cells @xcite and bacteria attached to nanoparticles @xcite .
however , biological organisms have limited material properties and computational speed . instead , we consider machines based on plausible extensions of currently demonstrated nanoscale electronics , sensors and motors @xcite and relying on directed assembly @xcite .
these components enable nonbiological robots that are stronger , faster and more flexibly programmed than is possible with biological organisms .
a major challenge for nanorobots arises from the physics of their microenvironments , which differ in several significant respects from today s larger robots .
first , the robots will often operate in fluids containing many moving objects , such as cells , dominated by viscous forces .
second , thermal noise is a significant source of sensor error and brownian motion limits the ability to follow precisely specified paths .
finally , power significantly constrains the robots @xcite , especially for long - term applications where robots may passively monitor for specific rare conditions ( e.g. , injury or infection ) and must respond rapidly when those conditions occur .
individual robots moving passively with the circulation can approach within a few cell diameters of most tissue cells of the body . to enable passing through even the smallest vessels ,
the robots must be at most a few microns in diameter .
this small size limits the capabilities of individual robots . for tasks requiring greater capabilities
, robots could form aggregates by using self - assembly protocols @xcite . for robots reaching tissues through the circulation ,
the simplest aggregates are formed on the inner wall of the vessel .
robots could also aggregate in tissue spaces outside small blood vessels by exiting capillaries via diapedesis @xcite , a process similar to that used by immune cells @xcite .
aggregates of robots in one location for an extended period of time could be useful in a variety of tasks .
for example , they could improve diagnosis by combining multiple measurements of chemicals @xcite . using these measurements
, the aggregate could give precise temporal and spatial control of drug release @xcite as an extension of an _ in vitro _ demonstration using dna computers @xcite .
using chemical signals , the robots could affect behavior of nearby tissue cells .
for such communication , molecules on the robot s surface could mimic existing signalling molecules to bind to receptors on the cell surface @xcite .
examples include activating nerve cells @xcite and initiating immune response @xcite , which could in turn amplify the actions of robots by recruiting cells to aid in the treatment .
such actions would be a small - scale analog of robots affecting self - organized behavior of groups of organisms @xcite .
aggregates could also monitor processes that take place over long periods of time , such as electrical activity ( e.g. , from nearby nerve cells ) , thereby extending capabilities of devices tethered to nanowires introduced through the circulatory system @xcite . in these cases ,
the robots will likely need to remain on station for tens of minutes to a few hours or even longer
. the aggregate itself could be part of the treatment by providing structural support , e.g. , in rapid response to injured blood vessels @xcite .
aggregates could perform precise microsurgery at the scale of individual cells , extending surgical capabilities of simpler nanoscale devices @xcite .
since biological processes often involve activities at molecular , cell , tissue and organ levels , such microsurgery could complement conventional surgery at larger scales .
for instance , a few millimeter - scale manipulators , built from micromachine ( mems ) technology , and a population of microscopic devices could act simultaneously at tissue and cellular size scales , e.g. , for nerve repair @xcite . for medical tasks of limited duration
, onboard fuel created during robot manufacture could suffice . otherwise , the robots need energy drawn from their environment , such as converting externally generated vibrations to electricity @xcite or chemical generators @xcite .
power and a coarse level of control can be combined by using an external source , e.g. , light , to activate chemicals in the fluid to power the machines in specific locations @xcite , similar to nanoparticle activation during photodynamic therapy @xcite , or by using localized thermal , acoustic or chemical demarcation @xcite .
this paper examines generating power for long - term robot activity from reacting glucose and oxygen , which are both available in the blood .
such a power source is analogous to bacteria - based fuel cells whose enzymes enable full oxidation of glucose @xcite .
we describe a computationally feasible model incorporating aspects of microenvironments with significant effect on robot performance but not previously considered in robot designs , e.g. , kinetic time constants determining how rapidly chemical concentrations adjust to robot operations . as a specific scenario , we focus on modest numbers of robots aggregated in capillaries .
a second question we consider is how the robots affect surrounding tissue .
locally , the robots compete for oxygen with the tissue and also physically block diffusion out of the capillary . robot power generation results in waste heat , which could locally heat the tissue
the robot oxygen consumption could also have longer range effects by depleting oxygen carried in passing red blood cells . in the remainder of this paper , we present a model of the key physical properties relevant to power generation for robots using oxygen and glucose in the blood plasma . using this model , we then evaluate the steady - state power generation capabilities of aggregated robots and how they influence surrounding tissue .
[ sect.model ] we consider microscopic robots using oxygen and glucose available in blood plasma as the robots power source .
this scenario involves fluid flow , chemical diffusion , power generation from reacting chemicals and waste heat production . except for the simplest geometries ,
behaviors must be computed numerically , e.g. , via the finite element method @xcite .
computational feasibility requires a choice between level of detail of modeling individual devices and the scale of the simulation , both in number of devices and physical size of the environment considered . for microscopic biological environments relevant for nanorobots ,
detailed physical properties may not be known or measurable with current technology , thereby limiting the level of detail possible to specify .
this section describes our model .
the simplifying approximations are similar to those used in biophysical models of microscopic environments , such as oxygen transport in small blood vessels with diffusion into surrounding tissue @xcite .
we focus on steady - state behavior indicating long - term robot performance when averaged over short - term changes in the local environment such as individual blood cells ( exclusively erythrocytes , not white cells or platelets unless noted otherwise ) passing the robots .
[ sect.geometry ] [ cols="^,^ " , ] fig .
[ fig.oxygen concentration ] shows the distribution of oxygen in the tissue and plasma in the vessel near the robots .
the robots reduce the local oxygen concentration far more than the surrounding tissue , as seen by comparing with the vessel without robots .
most of the extra oxygen used by the robots comes from the passing blood cells , which have about 100 times the oxygen concentration of the plasma . within the vessel with the robots , the concentration in the plasma is lowest in the fluid next to the robots .
downstream of the robots is a recovery region where the concentration increases a bit as cells respond to the abruptly lowered concentration near the robots . in the low demand scenario ,
the concentration in the vessel just downstream of the robots is somewhat lower than in the surrounding tissue .
thus in this region , the net movement of oxygen is from the tissue into the vessel , where the fluid motion transports the oxygen somewhat downstream before it diffuses back into the tissue . in effect
, part of the oxygen entering the vessel travels through the tissue around robots to the downstream section of the vessel , in contrast to the pattern without robots where oxygen is always moving from the vessel into the surrounding tissue .
the streamlines in fig .
[ fig.oxygen concentration ] show that the laminar flow speeds up as the fluid passes through the narrower vessel section where the robots are stationed .
oxygen concentration , in @xmath0 , along a radial cross section from the center of the vessel to the outer edge of the tissue region .
the cross section is in the middle of the modeled section of vessel and tissue , corresponding to a vertical line in the center of each plot of fig .
[ fig.oxygen concentration ] .
the gray area indicates the interior of the vessel .
the left and right plots correspond to the low and high demand scenarios of table [ table.scenarios ] . in each plot , the upper curve is for the vessel without robots and the lower curve is for the vessel containing the @xmath1 ringset with pumps . for comparison ,
the dashed curves are solutions to the krogh model @xcite corresponding to the vessel without robots.,title="fig : " ] oxygen concentration , in @xmath0 , along a radial cross section from the center of the vessel to the outer edge of the tissue region .
the cross section is in the middle of the modeled section of vessel and tissue , corresponding to a vertical line in the center of each plot of fig .
[ fig.oxygen concentration ] .
the gray area indicates the interior of the vessel .
the left and right plots correspond to the low and high demand scenarios of table [ table.scenarios ] . in each plot , the upper curve is for the vessel without robots and the lower curve is for the vessel containing the @xmath1 ringset with pumps . for comparison ,
the dashed curves are solutions to the krogh model @xcite corresponding to the vessel without robots.,title="fig : " ] fig .
[ fig.co2 section ] gives another view of how the robots affect the oxygen concentration in the surrounding tissue .
the concentration is zero at the robot surface facing into the vessel .
the robots decrease the oxygen concentration somewhat but do not affect tissue power generation much since the concentration remains well above the threshold where power generation drops significantly , i.e. , @xmath2 given in table [ table.parameters ] .
however , at large distances from the vessel in the high demand scenario oxygen concentration is low enough to significantly decrease tissue power production .
this low level of oxygen also occurs when there are no robots .
[ fig.co2 section ] includes comparison with the simpler krogh model of oxygen transport to tissue from vessels without robots @xcite .
the krogh model assumes constant power density in the tissue and no diffusion along the vessel direction in the tissue . for the low demand scenario ,
the krogh model results are close to those from our model .
however in the high demand case , the krogh model has oxygen concentration drop to zero about @xmath1 from the vessel , due to the unrealistic assumption of constant power use rather than the decrease in power use at low concentrations given by eq .
( [ eq.tissue reaction rate ] ) .
oxygen flux to the robots ranges from about @xmath3 to @xmath4 with the zero - concentration boundary condition .
estimates of pump capabilities are up to @xmath5 @xcite , which is more than 100 times the actual flux to the robots .
such pumps could thereby maintain the zero concentration boundary condition . at an energy use of @xmath6 @xcite
, the pumps would require about @xmath7 per robot to handle the incoming flux , slightly reducing the power benefit of the pumps .
however , much of this pumping energy may be recoverable by adding a generator using the subsequent expansion of the reaction products to their lower partial pressure outside the robot @xcite .
this section describes the steady - state power available to the robots according to our model in various scenarios .
we first discuss the average per robot power in the aggregate , for both high and low capacity cases , which also indicates the total power available to the aggregate as a whole .
we then show how the power is distributed among the robots , based on their location in the ringset .
finally , we illustrate the qualitative features of these results in a simpler , analytically - solvable model to identify key scaling relationships between robot design choices and power availability .
l|cc|cc|cc|cc|cc|cc robot power generation capacity & & + inlet concentration @xmath8 ( @xmath9 ) ' '' '' & & & + pressure gradient @xmath10 ( @xmath11 ) ' '' '' & & & & & & + tissue power demand @xmath12 ( @xmath13 ) ' '' '' & 4 & 60 & 4 & 60 & 4 & 60 & 4 & 60 & 4 & 60 & 4 & 60 + 10-micron ringset ( with pumps ) & 12 & 8 & 14 & 12 & * 17 * & 11 & 24 & * 18 * & * 17 * & 11 & 24 & * 18 * + 10-micron ringset ( free diffusion ) & 11 & 7 & 12 & 10 & * 15 * & 10 & 22 & * 16 * & * 6 * & 3 & 8 & * 6 * + 1-micron ring ( with pumps ) & 44 & 27 & 49 & 36 & * 69 * & 36 & 99 & * 58 * & * 69 * & 36 & 99 & * 58 * + 1-micron ring ( free diffusion ) & 31 & 19 & 34 & 25 & * 49 * & 25 & 71 & * 38 * & * 9 * & 4 & 12 & * 7 * + table [ table.behavior : power ] gives the average power generated using the available oxygen , per robot within the aggregate . as expected , robots receive more oxygen and hence can generate more power when inlet concentration is high , fluid speed is high or tissue power demand is low . in the first two cases , the flow brings oxygen through the vessel more quickly ; in the last case , surrounding tissue removes less oxygen .
the less than 2-fold decrease in robot power generation in the face of a larger 2.5-fold decrease in @xmath14 inlet concentration from the arterial to the venous end of the capillary shows that robots extract more oxygen from red cells than these cells would normally release while passing the length of the vessel .
thus robots get some of their oxygen as `` new oxygen '' rather than just taking it from what the tissues would normally get .
this is possible because in this case robots create steeper concentration gradients than the tissue does .
the 10-micron ringset with pumps produces about the same power in the low and high demand scenarios , consuming oxygen at @xmath15 . comparing the different aggregate sizes shows lower power generation , per robot , in the large aggregate compared to the small one
this arises from the competition among nearby robots for the oxygen .
nevertheless the larger aggregate , with ten times as many robots , generates several times as much power in aggregate as the smaller one .
this difference identifies a design choice for aggregation : larger aggregates have more total power available but less on a per robot basis .
robots using pumps generate only modestly more power than robots relying on diffusion alone in our high capacity design example ( section [ sect.robot power ] ) . in this case , for robots without pumps , the power generation site density , @xmath16 , is sufficiently large that oxygen molecules diffusing into the robot are mostly consumed by the power generators near the surface of the robot before they have a chance to diffuse back out of the robot .
for such robots , power generators far from the plasma - facing surface receive very little oxygen and hence do not add significantly to the robot power production .
pumps give higher benefit for isolated rings of robots than for tightly clustered aggregates .
although not evaluated in the axially symmetric model used here , pumps may be even more significant for a single isolated robot on the vessel wall .
such a robot would not be competing with any other robots for the available oxygen , though would still compete with nearby tissue .
the low capacity robots have only @xmath17 the maximum power generating capability of the high capacity robots discussed above .
nevertheless , each robot s maximum power is several times larger than the limit due to available oxygen .
thus pumps allow the robots to produce the same power as given in table [ table.behavior : power ] for the high capacity robots .
the pumps ensure the absorbed oxygen is completely used by the smaller number of reaction sites by increasing the concentration of oxygen within the robots so eq . (
[ eq.reaction rate ] ) gives the same power generation in spite of the smaller value of @xmath16 .
on the other hand , the smaller number of reaction sites is a significant limitation for robots without pumps . comparing with table [ table.behavior : power ] shows pumps improve the average power by factors of about 3 and 8 for the 10 and 1-micron ringsets , respectively . comparing with high capacity robots without pumps shows the factor of 50 reduction in reaction sites only reduces average power by factors of about 3 and 5 for the 10 and 1-micron ringsets , respectively .
thus the reaction sites in the low capacity scenario are used more effectively than in the high capacity robots : with a smaller number of sites , each site does not compete as much with nearby sites for the available oxygen .
much of the power in robots without pumps is generated near the plasma - facing surface , where oxygen concentration is largest . in our case ,
the power for the high capacity robots is generated primarily within @xmath18 of the robot surface .
this observation suggests that a design with power generating sites placed near this surface instead of uniformly throughout the robot volume , as we have assumed , could significantly improve power generation for robots without pumps .
for example , placing all the reaction sites uniformly within the @xmath17 of the robot volume nearest the surface would increase the local reaction site density in that volume by a factor of 50 . for low capacity robots , this placement would increase @xmath16 to the same value as the high capacity case , but only in a narrow volume , within @xmath19 of the surface with the robot geometry we use . elsewhere in the robot with this design @xmath20 . while we might expect this concentration to increase power significantly ,
in fact we find only a small increase ( e.g. , 12% for the 10-micron ringset in the low demand scenario ) . thus concentrating the reaction sites near
the plasma - facing surface does not offer much of a performance advantage .
[ sect.power : distribution ] steady - state power generation , in picowatts , for robots as a function of their position along the vessel wall , starting from those at the upstream end of the ringset ( position 1 ) and continuing to those at the downstream end ( position 10 ) . the charts , corresponding to the low and high demand scenarios of table [ table.scenarios ] , compare robots with pumps , absorbing all oxygen reaching them , with robots relying on free diffusion ( i.e. , without pumps ) , and high and low capacity robots .
for robots with pumps , power for high and low capacity cases are the same.,title="fig : " ] steady - state power generation , in picowatts , for robots as a function of their position along the vessel wall , starting from those at the upstream end of the ringset ( position 1 ) and continuing to those at the downstream end ( position 10 ) . the charts , corresponding to the low and high demand scenarios of table [ table.scenarios ] , compare robots with pumps , absorbing all oxygen reaching them , with robots relying on free diffusion ( i.e. , without pumps ) , and high and low capacity robots .
for robots with pumps , power for high and low capacity cases are the same.,title="fig : " ] while all robots in a single ring have the same power due to the assumption of axial symmetry , fig .
[ fig.power generation ] shows that power varies with ring position in the 10-micron ringset .
the robots at the upstream edge of the aggregate receive more oxygen than the other robots and hence produce more power .
power generation does not decrease monotonically along the vessel : robots at the downstream edge have somewhat more available oxygen than those in the middle of the aggregate since robots at the edge of the aggregate have less competition for oxygen .
[ fig.power generation ] shows significantly larger benefits of pumps at the edges of multi - ring aggregates than in their middle sections , especially in the low demand scenario . in the scenarios described above
, robots produce power from all the available oxygen .
this is appropriate for applications requiring as much power as possible for the aggregate as a whole . at the other extreme
, an application requiring the same behavior from all robots in the aggregate would be limited by the robots with the least available power .
this would be the case for identical robots , all of which perform the same task and hence use the same power . in this case , the robots could increase performance by transferring power from those at the edges of the aggregate to those in the middle .
such transfer could take place after generation , e.g. , via shared electric current , or prior to generation by transfer of oxygen among neighboring robots .
however , such internal transfer would require additional hardware capabilities . for robots with pumps ,
an alternative transfer method is for robots near the edge of the aggregate to run their pumps at lower capacity and thus avoid collecting all the oxygen arriving at their surfaces .
this uncollected oxygen would then be available for other robots , though some of this oxygen would be transported by the fluid past the robots or be captured by the tissue rather than other robots .
this approach increases power to robots in the middle of the aggregate without requiring additional hardware for internal transfers between robots , but at the cost of somewhat lower total power for the aggregate . increasing as much as possible the power to the robots with the least power leads to a uniform distribution of power among the robots . to quantify the trade - off between total power and its uniformity among the robots
, we consider all robots setting each of the pumps on their surfaces to operate at the same rate and the pumps uniformly distributed over the surface .
this gives a uniform flux of oxygen over the entire surface of all the robots .
the largest possible value for this uniform flux , and hence the largest power for the aggregate , occurs when the minimum oxygen concentration on the robot surfaces is zero at that point , the robot whose surface includes the location of zero concentration can not further increase its uniform flux . for example , in the low demand scenario the maximum value for this uniform flux is approximately @xmath21 .
compared to the situation in fig .
[ fig.power generation ] , this uniform flux gives significantly lower power ( 39% leading ring , 56% trailing ring ) for the robots at the edges of the aggregate , somewhat lower power for robots in positions 2 and 3 ( 87% and 98% , respectively ) and somewhat more power ( ranging from 104% to 116% ) for the other robots .
the combination of these changes gives a total of @xmath22 of the power for the aggregate when every robot collects all the oxygen reaching its surface .
the minimum power per robot increases from @xmath23 to @xmath24 .
robots could slightly increase power by accepting some nonuniformity of flux over each surface while maintaining the same total flux to each robot . this would occur when the minimum oxygen concentration on the entire length of the robot surfaces in a particular robot ring is zero . using this approach to uniform power in practice would require the robots to determine the maximum rate they can operate their pumps while achieving uniform power distribution .
this rate would vary with tissue demand , and also over time as cells pass the robots .
a simple control protocol is for each robot to adjust its pump rate up or down according to whether its power generation is below or above that of its neighbors , respectively .
when concentration reaches zero on one robot , increasing pump rate at that location would not increase power generation .
communicating information for this protocol is likely simpler than the hardware required to internally transfer power or oxygen among robots , but also requires each robot is able to measure its power generation rate .
such measurements and communication would give an effective control provided they operate rapidly compared to the time over which oxygen flux changes , e.g. , as cells pass the robots on millisecond time scales .
longer reaction times could lead to oscillations or chaotic behavior @xcite .
a second approach to achieving a more uniform distribution of power is to space the robots at some distance from each other on the vessel wall .
this approach would be suitable if the aggregated robots do not need physical contact to achieve their task .
for example , somewhat separating the robots would allow a relatively small number to span a distance along the vessel wall larger than the size of a single cell passing through the vessel .
this aggregate would always have at least some robots between successive cells .
communicating sensor readings among the robots would then ensure the response , e.g. , releasing chemicals , is not affected by misleading sensor values due to the passage of a single cell , giving greater stability and reliability without the need for delaying response due to averaging over sensor readings as an alternative approach to accounting for passing cells .
another example for spaced robots is for directional acoustic communication , at distances of about @xmath25 @xcite .
achieving directional control requires acoustic sources extending over distances comparable to or larger than the sound wavelength .
plausible acoustic communication between nanorobots involves wavelengths of tens of microns @xcite . as a quantitative example of the benefit of spacing robots in the context of our axially symmetric model
, we consider a set of rings of robots spaced apart along the vessel wall . when the distance between successive rings is sufficiently large , the power for each ring would be close to that of the isolated 1-micron ring given in table [ table.behavior : power ] .
for example , the power for the low demand scenario in the 1-micron ring of high - capacity robots decreases from @xmath26 at high inlet concentration to @xmath27 at low inlet concentration , which spans the range of power for a modest number of widely spaced 1-micron rings within a single vessel .
as described in section [ sect.tissue power and heating ] , oxygen absorption by robots can affect concentration over a few tens of microns upstream of those robots .
thus separating robot rings by , say , @xmath25 will give power close to that of the isolated rings , with a gradual decrease in power for successive rings due to the decreasing cell saturation along the vessel . a third approach to reducing variation in robot power , on average ,
is through changing pump rates in time .
for example , adjacent nanorobot rings could operate with counterphased 50% duty cycles , with one ring and its second nearest neighbor ring using pumps while the intervening nearest neighbor has its pumps off and does not absorb oxygen .
the alternating rings of robots would switch pumps on and off . in this case
, robots would have larger power than seen in fig . [ fig.power generation ] for the half of the time they are active , and zero power for the other half .
this temporal approach would not be suitable for tasks requiring all robots to have the same power _ simultaneously _ , but would be useful for tasks requiring higher burst power from robots throughout the aggregate where the robots are unable to store oxygen or power for later use .
provided the duty cycle is sufficiently long , our steady - state model can quantify the resulting power distribution .
for example , in the low demand scenario , total flux for the aggregate is 79% of that when every robot collects all the oxygen reaching its surface , and the minimum power per robot _ drops _ from @xmath23 to @xmath28 .
thus , when averaged over the duty cycle , this temporal technique reduces total power without benefiting the robots receiving the minimum power . in this case
, the temporal approach does not improve minimum robot power ( on average ) since the power gain to a robot while its neighbors are off is less than a factor of 2 , which does not compensate for the loss due to each robot being off for half the time . applying the steady - state model to this temporal variation in robot activity requires the duty cycle be long enough for the system to reach steady - state behavior after each switch between the active subset of robots , and that the switching time is short compared to the duty cycle so most of the robots power arises during the steady - state portions of the cycle between switching .
diffusion provides one lower bound on this time : when neighboring robots switch pumps from on to off or vice versa , the characteristic diffusion time for oxygen over the distance between next nearest neighbors ( one micron ) is about @xmath29 .
adjustments in cell saturation for the 1-micron shift in the location of the active robots between each half of the duty cycle is a further limitation on the duty cycle time for the validity of the steady - state model , though this is likely to be minimal since the cells are separated from the robots by the plasma gap in the fluid . since the steady - state model averages over the position of passing cells , another lower bound on the duty cycle arises from the time for a cell to pass the robots . from the speeds in table [ table.parameters ] ,
this time is at least @xmath30 .
the dependence of robot power on design parameters described above may appear contrary to simple intuitions .
first , one might expect that the 10-micron ringset , with ten times the surface area in contact with the plasma , would absorb about ten times as much oxygen as the 1-micron ring .
instead we find only about a factor of 2 to 4 increase .
second , the benefit of pumps , less than a factor of 2 for the high capacity robots , may seem surprisingly small .
third , the low capacity robots , with @xmath17 the reaction sites of the high capacity robots , nevertheless generate about @xmath31 as much power as high capacity robots in the case with no pumps . and
finally , in spite of the higher concentration near the robot surface than deep inside the robot when there are no pumps , increasing the reaction site density by placing all the reaction sites near the robot surface gives little benefit .
while the specific values of these designs depend on the geometry and environment used in our model , these general features of small robots obtaining power through diffusion apply in other situations as well . in this section
we illustrate how these consequences of design choices arise in the context of a scenario for which the diffusion equation has a simple analytic solution , thereby identifying key physical effects leading to these behaviors . specifically
, we consider an isolated spherical robot of radius @xmath32 in a stationary fluid with oxygen concentration @xmath33 far from the sphere .
such a sphere with a fully absorbing surface collects oxygen at a rate @xmath34 @xcite .
this expression illustrates a key property of diffusive capture : the rate depends not on the object s surface area but on its size .
this behavior , which also applies to other shapes @xcite , arises because while larger objects have greater surface areas they also encounter smaller concentration gradients . as a quantitative example , taking the sphere to have the same volume as the robot , i.e. , @xmath35 given in table [ table.robot parameters ] , the oxygen absorbed by the sphere generates @xmath36 and @xmath37 for @xmath33 equal to the low and high inlet oxygen concentrations in the plasma from table [ table.parameters ] , respectively . these power values are larger than for robots on the vessel wall described above .
unlike the sphere in a stationary fluid , the aggregated robots compete with each other for the oxygen , the fluid moves some of the oxygen past the robots before they have a chance to absorb it , and the surrounding tissue also consumes some of the oxygen .
the replenishment of oxygen from the passing blood cells is not sufficient to counterbalance these effects .
the spherical robot also indicates the benefit of pumps .
the fully absorbing sphere , with a zero concentration boundary condition at the surface , corresponds to using pumps . for robots without pumps ,
an approximation to eq .
( [ eq.reaction rate ] ) allows a simple solution .
specifically , since the michaelis - menten constant for the robot power generators , @xmath38 , is much larger than the oxygen concentrations ( e.g. , as seen in fig . [ fig.oxygen concentration ] ) , robot power generation from eq .
( [ eq.reaction rate ] ) is approximately @xmath39 .
dividing by @xmath40 gives the oxygen consumption rate density as @xmath41 where @xmath42 .
solving the diffusion equation , eq .
( [ eq.diffusion ] ) , for a sphere in a stationary fluid with concentration @xmath33 far from the sphere , with free diffusion through the sphere s surface and reaction rate density @xmath41 inside the sphere , gives the rate oxygen is absorbed by the sphere ( and hence reacted to produce power ) as @xcite @xmath43 where @xmath44 .
thus free diffusion produces a fraction @xmath45 of the power produced by the fully absorbing sphere .
the distance @xmath46 is roughly the average distance an oxygen molecule diffuses in the time a power generation site consumes an oxygen molecule .
when a freely diffusing molecule inside the robot has a high chance to diffuse out of the robot before it reacts ( @xmath46 large compared to @xmath32 ) , @xmath47 is small so pumps provide a significant increase in power .
conversely , when @xmath46 is small compared to @xmath32 , pumps provide little benefit : the large number of reaction sites ensure the robot consumes almost all the diffusing oxygen reaching its surface .
this argument illustrates a tradeoff between using pumps to keep oxygen within the robot and the number of power generators .
in particular , if internal reaction sites are easy to implement , then robots with many reaction sites and no pumps would be a reasonable design choice .
conversely , if reaction sites are difficult to implement while pumps are easy , then robots with pumps and few reaction sites would be a better choice .
a caveat for robots with few power generating sites is that eq .
( [ eq.f mu ] ) applies when oxygen consumption is linear in the concentration , as given by @xmath41 .
this expression allows arbitrarily increasing the reaction rate by increasing the concentration , no matter how small the number of reaction sites .
this linearity is a good approximation of eq .
( [ eq.reaction rate ] ) only when @xmath48 . at larger concentrations the power density saturates at @xmath49 .
when @xmath16 is sufficiently small , this limit is below the power that could be produced from all the oxygen that a fully absorbing sphere collects .
thus , in practice , the benefit of using pumps estimated from the linear reaction rate , @xmath50 , is limited by this bound when @xmath16 is small . as an example , for a spherical robot with the high capacity reaction site density of table [ table.robot parameters ] , @xmath51 and @xmath52 , with the fairly modest benefit of pumps @xmath53 .
the low capacity robots have @xmath54 and @xmath55 with @xmath56 . in this case , the limit due to the maximum reaction rate of eq . (
[ eq.reaction rate ] ) applies , somewhat limiting the benefit of pumps to a factor of @xmath57 , but pumps still offer considerable benefit .
these values for the benefits of pumps are somewhat larger than seen with our model for robots on the vessel wall .
nevertheless , the spherical example identifies the key physical properties influencing power generation with and without pumps , and how they vary with robot design choices .
( [ eq.f mu ] ) also illustrates why power in the low capacity robots is not as small as one might expect based on the reduction in reaction sites by a factor of 50 .
while the value of @xmath58 is proportional to @xmath59 , the typical diffusion distance @xmath46 varies as @xmath60 , so a decrease in reaction sites by a factor of 50 only increases @xmath46 by about a factor of 7 .
the square root dependence arises from the fundamental property of diffusion : typical distance a diffusing particle travels grows only with the square root of the time .
the modest change in diffusion distance , combined with eq .
( [ eq.f mu ] ) , gives a smaller decrease in power than the factor of 50 decrease in capacity .
the low capacity robot has higher concentrations throughout the sphere , so each reaction site operates more rapidly than in the high capacity case .
this increase partially offsets the decrease in the number of reaction sites . without pumps , the higher oxygen concentration near the sphere s surface than near
its center means much of the power generation takes place close to the surface .
thus we can expect an increase in power by placing reaction sites close to the surface rather than uniformly distributed throughout the sphere .
consistent with the results from the model described in section [ sect.model ] , evaluating eq .
( [ eq.diffusion ] ) with the reaction confined to a spherical shell shows only a modest benefit compared to a uniform distribution . the benefit is larger for a thinner shell and
is determined by the same ratio , @xmath61 , appearing in eq .
( [ eq.f mu ] ) .
in particular , the largest benefit of using a thin shell , only @xmath62 , occurs for @xmath63 .
the parameters for the high and low capacity robots are somewhat below this optimal value , giving only @xmath64 and less than @xmath65 benefit from a thin shell , respectively , for the sphere .
these modest improvements correspond to the small benefits of using a thin shell seen in the solution to our model for both high and low capacity robots .
hence the solution of eq .
( [ eq.diffusion ] ) for the sphere illustrates how , with a fixed number of reaction sites , concentrating them near the robot surface provides only limited benefit .
the benefit of the higher reaction site density in the shell is almost entirely offset by the shorter distance molecules need to diffuse to escape from the thin reactive region .
that is , the benefit of placing all the reaction sites in a thin shell arises from two competing effects .
when @xmath46 is large ( low capacity ) the concentration is only slightly higher near the surface than well inside the sphere .
so there is little benefit from placing the reaction sites closer to the surface . on the other hand ,
when @xmath46 is small ( high capacity ) , even uniformly distributed reaction sites manage to consume most of the arriving oxygen , giving near - zero concentration at the surface of the sphere and little scope for further improvement by concentrating the reaction sites .
thus the largest , though still modest , benefit for a shell design is for intermediate values of @xmath61 .
deviation from equilibrium oxygen saturation in cells , @xmath66 , along the boundary between the cell and cell - free portions of the fluid illustrated in fig .
[ fig.geometry ] , as a function of distance along that boundary .
a deviation of zero indicates the oxygen held in the cells is in perfect equilibrium with the surrounding plasma .
saturation ranges between 0 and 1 .
the curves correspond to the low and high demand scenarios of table [ table.scenarios ] , when robots are present ( upper curves ) or absent ( lower curves ) . for the vessel without robots and low demand ,
@xmath66 is indistinguishable from zero on the scale of the plot .
the gray band indicates the 10-micron length of the vessel wall in which the robots are stationed , within the 60-micron length of the capillary illustrated . ]
the high power density of the robots creates a steep gradient of oxygen concentration in the plasma .
thus , unlike the minor role for nonequilibrium oxygen release in tissue @xcite , the small size of the robots makes passing red cells vary significantly from equilibrium with the concentration in the plasma .
[ fig.s - sequib ] illustrates this behavior , using one measure of the amount of disequilibration : the difference between saturation @xmath67 and the equilibrium value @xmath68 corresponding to the local concentration of oxygen in the plasma , as given by eq .
( [ eq.equilibrium saturation ] ) .
we compare with a vessel without robots , in which the blood cells remain close to equilibrium .
[ fig.s - sequib ] shows that the kinetics of oxygen release from red cells plays an important role in limiting the oxygen available to the robots .
however , the region of significant disequilibration is fairly small , extending only a few microns from the robots .
[ sect.tissue power and heating ] power density in tissue next to the vessel wall relative to maximum demand , i.e. , the ratio @xmath69 from eq .
( [ eq.tissue reaction rate ] ) , as a function of position along the vessel .
the curves are for the low and high demand scenarios of table [ table.scenarios ] .
the two lines at the top are for the vessel without robots and the lower curves are for the 10-micron ringsets . in each case
, the curve with higher values corresponds to the low power demand scenario .
the gray band indicates the 10-micron length of the vessel wall in which the robots are stationed , within the 60-micron length of the capillary illustrated . ]
cell saturation @xmath67 as a function of distance along the vessel .
saturation ranges between 0 and 1 .
the curves correspond to the low and high demand scenarios of table [ table.scenarios ] , with the upper curves of each pair corresponding to a vessel without robots .
the gray band indicates the 10-micron length of the vessel wall in which the robots are stationed , within the 60-micron length of the capillary illustrated . ]
the robots affect tissue power in two ways .
first , the robots compete for oxygen with nearby tissue .
second , the robots consume oxygen from passing blood cells , thereby leaving less for tissue downstream of the robots . for the effect on nearby tissue , fig .
[ fig.tissue power ] shows how tissue power density varies next to the vessel wall . in the vessel without robots ,
power density declines slightly with distance along the vessel as the tissue consumes oxygen from the blood .
the total reduction in tissue power density is fairly modest , less than 10% even for high power demand in the tissues . the relative reduction is less for tissue at larger distances from the vessel , though such tissue has lower power generation due to less oxygen reaching tissue far from the vessel .
this reduction arises both from direct competition by the robots for available oxygen and the physical blockage of the capillary wall , forcing surrounding tissue to rely on oxygen diffusing a longer distance from unblocked sections of the wall . in the low demand case , direct competition is the major factor , as seen by the dips in the power density at each end of the aggregate , where the absorbing flux is highest . in the high demand case ,
the tissue s consumption reduces the amount of oxygen diffusing through the tissue on either side of the aggregate , giving the larger drop in tissue power density in the middle of the aggregate . for longer range consequences ,
[ fig.saturation ] shows how the oxygen saturation in the blood cells changes as they pass the robots . slowly moving cells ( in the low demand scenario )
are substantially depleted while passing the robots , even though tissue power demand in this scenario is low .
this depletion arises from the cells remaining near the robots a relatively long time as cells move slowly with the fluid .
the resulting saturation shown in the figure , around @xmath70 , is below the equilibrium saturation ( @xmath71 ) for typical concentrations at the venous end of capillaries , given in table [ table.parameters ] .
thus in the low demand scenario , the robots remove more oxygen from passing cells than occurs during their full transit of a vessel containing no robots . in this scenario ,
the tissue has low power demand , so the depletion of oxygen from the cells may have limited effect on tissue along the vessel downstream of the robots .
however , this reduction could significantly limit the number of robots that can be simultaneously present inside a given capillary .
another observation from fig .
[ fig.saturation ] is a significant decrease in cell saturation a short distance _ upstream _ of the robots in the low demand scenario
. we can understand this behavior in terms of the peclet number , which characterizes the relative importance of convection and diffusion over various distances @xcite .
in particular , @xmath72 is the distance at which diffusion and convection have about the same effect on mass transport in a moving fluid . at significantly longer distances , convection is the dominant effect and absorption of oxygen at a given location in the vessel has little effect on upstream concentrations over such distances . in our scenarios , @xmath72 ranges from @xmath73 ( high demand ) to @xmath74 ( low demand ) .
thus the oxygen concentration in the plasma is significantly affected by the robots over a few tens of microns upstream of their location .
cell saturation remains close to equilibrium in this upstream region ( fig .
[ fig.s - sequib ] ) , hence the reduced oxygen concentration in the plasma lowers cell saturation in this region upstream of the robots ( fig .
[ fig.saturation ] ) .
this distance is also relevant for spacing rings of robots far enough apart to achieve nearly uniform power , as described in section [ sect.power : distribution ] .
the devices in this example have a volume of @xmath75 so the robot power generation corresponds to power densities around @xmath76 , several orders of magnitude larger than power densities in tissue , raising concerns of possible significant tissue heating by the robots .
however , for the isolated aggregate used in this scenario , waste heat due to the robots power generation is rapidly removed , resulting in negligible maximum temperature elevation of about @xmath77 .
the robots change the fluid flow by constricting the vessel . with the same pressure difference as a vessel without robots , as used in our model , this constriction results in somewhat lower flow speed through the vessel . specifically , the one and ten - micron long aggregates reduce flow speed by 6% and 20% , respectively .
the fluid moving past the robots exerts a force on them . to remain on the wall
the robots must resist this force through their attachment to the vessel wall @xcite .
this force is a combination of pressure difference , between the upstream and downstream ends of the aggregate , and viscous drag . for the laminar flow the force
@xmath78 is linear in the pressure gradient @xmath10 imposed on the vessel : @xmath79 where @xmath80 and @xmath81 for the @xmath1 and @xmath82 ringsets , respectively .
for example , the flow imposes a force of @xmath83 on the @xmath1 ringset when the pressure gradient is @xmath11 .
the @xmath1 ringset experiences about three times the force of the @xmath82 ring , but covers ten times the surface area .
thus the larger aggregate requires about one - third the attachment force per robot .
applied forces can affect cells @xcite .
in particular , endothelial cells use forces as a trigger for new vessel growth @xcite , which is important for modeling changes in the vessels over longer time scales than we consider in this paper @xcite .
[ sect.discussion ] the scenarios of this paper illustrate how various physical properties affect robot power generation . robots about one micron in size positioned in rings on capillary walls could generate a few tens of picowatts in steady state from oxygen and glucose scavenged locally from the bloodstream .
aggregates can combine their oxygen intake for tasks requiring higher sustained power generation .
the resulting high power densities do not significantly heat the surrounding tissue , but do introduce steep gradients in oxygen concentration due to the relatively slow reaction kinetics of oxygen release from red cells .
the robots reduce oxygen concentration in nearby tissues , but generally not enough to significantly affect tissue power generation .
the fraction of the generated power available for useful activity within the robot depends on the efficiency of the glucose engine design , with @xmath84 a reasonable estimate for fuel cells @xcite .
the robots will have @xmath85 of usable steady - state power while on the vessel wall .
as one indication of the usefulness of this power for computation , current nanoscale electronics and sensors have an energy cost per logic operation or sensor switching event of a few hundred @xmath86 @xcite .
while future technology should enable lower energy use @xcite , even with @xmath87 the available power from circulating oxygen could support several million computational operations per second . at the size of these robots , significant movements of blood cells and
chemical transport occur on millisecond time scales .
thus the power could support thousands of computational operations , e.g. , for chemical pattern recognition , in this time frame .
the aggregated robots could share sensor information and cpu cycles , thereby increasing this capability by a factor of tens to hundreds .
the robots need not generate power as fast as they receive oxygen , but could instead store oxygen received over time to enable bursts of activity as they detect events of interest
. robots with pumps have a significant advantage in burst - power applications because pumps enable long - term high - concentration onboard gas accumulation to support brief periods of near maximal power generation . as an example , in our scenarios individual robots with pumps receive about @xmath88 .
if instead of using this oxygen for immediate power generation , the robot stored the oxygen received over one second , it would have enough to run the power generators at near maximal rate ( giving about @xmath89 ) for several milliseconds .
by contrast , robots without pumps would only have a modest benefit from oxygen diffusing into the robot , achieving a concentration equal to the ambient concentration in the surrounding plasma as given in table [ table.parameters ] . this concentration could only support generating several hundred picowatts for about a tenth of a millisecond . thus while pumps may give only modest improvement for steady - state power generation , they can significantly increase power available in short bursts . our model could be extended to estimate the amount of onboard storage that would be required to avoid pathological conditions related to @xmath14 competition between tissues and nanorobots . in particular , larger aggregates would deplete oxygen over longer distances for which diffusion through the tissue from upstream of the robots would be insufficient . furthermore
, larger aggregates of tightly - spaced robots would block transport from the capillary into the surrounding tissue even if the robots did not use much oxygen .
onboard oxygen storage would allow higher transient power densities for the robots , though this could lead to heating issues for larger aggregates .
to estimate the potential for onboard storage , a ringset containing 200 robots with volume @xmath90 of which 10% is devoted to compressed @xmath14 storage at 1000 atmospheres at body temperature can store about @xmath91 molecules of @xmath14 in the aggregate .
the incoming flow in the capillary provides about @xmath92 molecules per second , depending on the flow speed , of which about @xmath93 to @xmath94 is available to the tissue and robots .
this means the oxygen stored in the aggregate is equivalent to only several seconds of oxygen delivery through the vessel .
thus oxygen storage in the robots themselves can not significantly increase mission duration , though such storage might be useful for short - term ( i.e. , a few seconds ) load leveling functions ( e.g. , maintaining function during temporary capillary blockage due to white cell passage ) .
alternatively , the aggregated robots could have oxygen supplemented with a modest circulating population of respirocytes @xcite , i.e. , @xmath82 spherical robots able to carry oxygen to tissues far more effectively than red blood cells . such robots would continuously and entirely eliminate any oxygen depletion regions in the tissue due to robot power generation , and allow higher robot power generation since oxygen would no longer be such a limiting factor .
such machines could not only carry significantly more oxygen than red blood cells , but would also respond more quickly to abrupt decreases in partial pressure due to consumption by aggregated robots on vessel walls .
for example , sensors should be able to detect the drop in concentration of the size we see near the robots
e.g. , @xmath95 ( or about @xmath96 ) within a millisecond @xcite .
this time is short enough that the machines will have moved only about a micron and so will still be near the robots .
once they detect the pressure drop , the machines could release oxygen rapidly , up to @xmath97 @xcite , while passing near the aggregated robots .
however , in practice the release rate is constrained by the effervescence limit in plasma to about @xmath98 @xcite .
an interesting question for future work is evaluating how much of this released oxygen reaches the robots on the vessel wall , which will depend on how close to the vessel wall the fluid places the respirocytes .
in this situation , aggregated robots could also communicate @xcite to passing respirocytes to activate or suppress their oxygen delivery , depending on the task at hand .
thus both the oxygen handling capabilities of respirocytes , giving faster kinetics than red cells and larger storage capacity , and the possibility of communication provide examples of the flexibility of small devices with programmable control .
moreover , this scenario illustrates the benefits of mixing robots with differing hardware capabilities .
our model considers static aggregates on vessel walls , but could be extended to study power availability for aggregates that move along the walls @xcite .
another significant scenario is robots moving passively with the fluid , where they could draw oxygen from the surrounding plasma .
the oxygen unloading model used here could evaluate how rapidly nearby cells would replenish oxygen in the plasma as the cells and robots move through the capillary .
we treat environmental parameters ( e.g. , fluid flow speed and tissue oxygen demand ) as fixed by the surroundings . beyond the local changes in the robots environment described by the model , sustained use of these robots could induce larger scale responses .
for example , the increased use of oxygen by the robots could lead to increased blood flow , as occurs with , say , exercising muscles , by increased pressure to drive the fluid at higher speed or dilation of the vessels .
the local oxygen deficits due to high robot power use are smaller in scale than higher tissue demand ( e.g. , from increased activity in a muscle ) .
thus an important open question is whether localized robot oxygen consumption over a long period of time can initiate a less localized response to increase flow in the vessels .
the possibility of large - scale responses to robot activity raises a broader issue for nanomedicine treatment design when technology allows altering the normal correlation among physical quantities used for signalling in the body @xcite .
an example at larger scales is the response to low oxygen mediated by excess carbon dioxide in the blood , which can lead to edema and other difficulties for people at high altitudes @xcite . in terms of downstream consequences of the robots oxygen use
, low saturation of cells leaving an isolated capillary should not be a problem because the bulk of oxygen exchange occurs in the capillary bed , not in the larger collecting vessels .
however , cells reaching low saturation before exiting the capillary would produce localized anoxia in the tissue near the end of the capillary .
this could be relieved in part by oxygen diffusion from neighboring tissue cells if the anoxic region is not too large or too severe .
specific effects of such localized anoxia remain to be fully identified .
whole capillaries subjected to ischemic conditions over a period of days remodel themselves , e.g. , by adding new vascular branches and by increasing the tortuosity of existing vessels @xcite .
this observed behavior is likely to be a localized ( i.e. , cell - level ) response , hence we might expect such a response if a portion of a capillary downstream of the robots was driven into ischemic conditions .
there could also be a localized inflammatory response to a large enough number of capillary - wall endothelial cells under stress , especially for cells stressed to the point of apoptosis , but moderate ischemia alone seems unlikely to generate this response @xcite .
various chemicals ( e.g. , adrenalin ) make the heart pump faster and thus drive the blood at higher speed .
other chemicals ( such as no , pgd2 and ltd4 ) dilate the vessels .
these chemicals can produce significant activity in the endothelial cells that line ( and thus form the tube geometry of ) the capillary vessel , so their influence can be fairly direct and quick @xcite .
similarly , a large robot population constantly drawing excess oxygen supply could induce elevated erythropoietin secretion ( if unregulated by the robots ) , increasing red cell production in the erythroid marrow @xcite .
direct heating is not a problem with aggregates of the size considered here , in spite of their high power density compared to tissue . for the large aggregate we examined ( tightly covering @xmath1 along the vessel wall ) , oxygen diffusion through the tissue from regions upstream and downstream of the robots provided oxygen to the tissue outside the section of the vessel blocked by the robots .
larger aggregates , especially if tightly packed , would significantly reduce oxygen in the tissues even if the robots used little power themselves , simply due to their covering the vessel wall over a long enough distance that diffusion through the tissue from unblocked regions is no longer effective .
the inducement of nonlinear tissue thermal responses ( e.g. , inflammation or fever ) due to the heat generated by larger aggregates or multiple aggregates in nearby capillaries is an important question for future work .
nanorobots parked or crawling along the luminal surface of the vessel may activate mechanosensory responses from the endothelial cells across whose surfaces the nanorobots touch @xcite .
if the aggregates cover a long section of the vessel wall , they could produce local edemas since narrowing of the vessels by the presence of the nanorobots increases local pressure gradients and fluid velocities .
while we focus on a single aggregate in one microscopic vessel , additional issues arise if a large population of circulating robots form many aggregates .
in that case , the multiple aggregates will increase hydrodynamic resistance throughout the fluidic circuit .
thus the robots could make the heart work slightly harder to pump fluid against the slightly higher load .
moreover , if robot aggregates detach from the wall without complete disaggregation , these smaller aggregates moving in the blood may be large enough to block a small vessel .
the scenarios examined in this paper can suggest suitable controls to distribute power when robots aggregate . moreover , power control decisions interact with the choices made for the aggregation process .
for example , if the task requires a certain amount of total power for the aggregate ( e.g. , as a computation hub ) then the aggregation self - assembly protocol would depend on how much oxygen is available , e.g. , to make a larger aggregate in vessels with less available oxygen , or recruit more passing robots when the task needs more power .
an example of this latter case could be if aggregates are used as computation hubs to validate responses to rare events : when local sensor readings indicate the possibility of such an event , the aggregate could temporarily recruit additional robots to increase power and computational capability for evaluating whether those readings warrant initiating treatment .
another approach to designing controls for teams of robots is the formalism of partially observable markov processes @xcite .
this formalism allows for arbitrarily complex computations among the robots to update their beliefs about their environment and other robots .
unfortunately this generality leads to intractable computations for determining optimal control processes .
for the situations we studied , the power constraints on capillary wall - resident microscopic robots operating with oxygen available _ in vivo _ means the local rules must be simple . including this constraint in the formalism could allow it to identify feasible control choices for large aggregates of microscopic robots in these situations .
the power constraints from our model could provide useful parameters for less detailed models of the behavior of large numbers of robots in the circulation in the context of the scenarios examined in this paper .
in particular , power limits the computation , communication and locomotion capabilities of the robots .
these constraints could be incorporated in simplified models , such as cellular automata approaches to robot behavior .
these automata are a set of simple machines , typically arranged on a regular lattice .
each machine is capable of communicating with its neighbors on the lattice and updates its internal state based on a simple rule .
for example , a two - dimensional scenario shows how robots could assemble structures @xcite using local rules .
such models can help understand structures formed at various scales through simple local rules and some random motions @xcite .
a related analysis technique considers swarms @xcite , i.e. , groups of many simple machines or biological organisms such as ants . in these systems , individuals use simple rules to determine their behavior from information about a limited portion of their environment and neighboring individuals .
typically , individuals in swarms are not constrained to have a fixed set of neighbors but instead continually change their neighbors as they move .
swarm models are well - suited to microscopic robots with their limited physical and computational capabilities and large numbers .
most swarm studies focus on macroscopic robots or behaviors in abstract spaces @xcite . in spite of the simplified physics ,
these studies show how local interactions among robots lead to various collective behaviors @xcite and provide broad design guidelines . a step toward more realistic , though still tractable
, models of large aggregates could incorporate the power constraints from the model presented in this paper .
in addition to evaluating performance of hypothetical medical nanorobots , theoretical studies identifying tradeoffs between control complexity and hardware capabilities can aid future fabrication .
one example is the design complexity of the robot s fuel acquisition and utilization systems . for steady - state operation on vessel walls , we found limited benefit of pumps over free diffusion when numerous onboard power generators can be employed . in such cases ,
our results indicate that a design without pumps does not sacrifice much performance .
more generally , control can compensate for limited hardware ( e.g. , sensor errors or power limitations ) , providing design freedom to simplify the hardware through additional control programs .
thus the studies could help determine minimum hardware performance capabilities needed to provide robust systems - level behavior .
a key challenge for robot design studies based on approximate models is validating the results . in our case ,
the most significant approximations are the treatment of cells as an averaged component in the fluid and the lumped - model kinetics for oxygen unloading . with increased computational power , numerical solution of more accurate models
could test the validity of these approximations . as technology
advances to constructing early versions of microscopic robots , experimental evaluations will supplement theoretical studies .
one such experiment is operating the robots in manufactured microfluidic channels @xcite .
this would test the robots ability to aggregate at chemically defined locations and generate power reliably from known chemical concentrations in the fluid .
after such _ in vitro _
experiments , early _ in vivo _ tests could involve robots acting as passive sensors in the circulatory system and aggregating at chemically distinctive locations
. such nanorobots will be useful not only as diagnostic tools and sophisticated extensions to drug delivery capabilities @xcite , but also as an aid to develop robot designs and control methods for more active tasks .
raf acknowledges private grant support for this work from the life extension foundation and the institute for molecular manufacturing .
jie bao , keiji furumoto , kimitoshi fukunaga , and katsumi nakao .
a kinetic study on air oxidation of glucose catalyzed by immobilized glucose oxidase for production of calcium gluconate .
, 8:91102 , 2001 .
anthony r. cassandra , leslie pack kaelbling , and michael l. littman .
acting optimally in partially observable stochastic domains . in
_ proc . of the 12th natl .
conf . on artificial intelligence
( aaai94 ) _ , pages 10231028 , menlo park , ca , 1994 .
aaai press .
heiko hamann , heinz worn , karl crailsheim , and thomas schmickl .
spatial macroscopic models of a bio - inspired robotic swarm algorithm . in _ proc .
conf . on intelligent robots and systems ( iros 2008)_. inria , 2008 .
tad hogg and bernardo a. huberman .
dynamics of large autonomous computational systems . in kagan tumer and david wolpert , editors , _ collectives and the design of complex systems _ , pages 295315 .
springer , new york , 2004 .
balazs l. keszler , istvan j. majoros , and james r. baker jr .
molecular engineering in nanotechnology : structure and composition of multifunctional devices for medical application . in _ proc . of the ninth foresight conference on molecular nanotechnology _ , 2001 .
scott p. leary , charles y. liu , and michael l. j. apuzzo . toward the emergence of nanoneurosurgery : part iii - nanomedicine : targeted nanotherapy , nanosurgery , and progress toward the realization of nanoneurosurgery .
, 58:10091026 , 2006 .
moustafa malki , antonio l. de lacey , nuria rodrguez , ricardo amils , and victor m. fernandez .
preferential use of an anode as an electron acceptor by an acidophilic bacterium in the presence of oxygen .
, 74:44724476 , 2008 .
dominik szczerba , gabor szekely , and haymo kurz . a multiphysics model of capillary growth and remodeling .
in v. n. alexandrov et al . , editors , _ proc . of iccs , part ii _ , pages 8693 , berlin , 2006 . springer . | the power available to microscopic robots ( nanorobots ) that oxidize bloodstream glucose while aggregated in circumferential rings on capillary walls is evaluated with a numerical model using axial symmetry and time - averaged release of oxygen from passing red blood cells .
robots about one micron in size can produce up to several tens of picowatts , in steady - state , if they fully use oxygen reaching their surface from the blood plasma .
robots with pumps and tanks for onboard oxygen storage could collect oxygen to support burst power demands two to three orders of magnitude larger .
we evaluate effects of oxygen depletion and local heating on surrounding tissue .
these results give the power constraints when robots rely entirely on ambient available oxygen and identify aspects of the robot design significantly affecting available power .
more generally , our numerical model provides an approach to evaluating robot design choices for nanomedicine treatments in and near capillaries .
* keywords : * nanomedicine , nanorobotics , capillary , power , numerical model , oxygen transport |
ferro ( ferri ) magnetic insulators ( fmi ) with normal metallic ( nm ) contacts that support the spin hall effect ( she ) and its inverse ( ishe ) open new functionalities in the field of spintronics .
the she refers to a charge current that induces a transverse spin current , which can be injected to actuate a metallic or insulating ferromagnet .
the ishe converts a spin current pumped out of a ferromagnet into a transverse charge current in the normal metal .
these concepts have been confirmed by many experiments on fmi@xmath5 nm bilayers of a magnetic insulator ( usually yttrium iron garnet ) and a heavy normal metal ( usually platinum ) , for example spin - pumping by ferromagnetic resonance @xcite , the spin seebeck effect ( sse ) @xcite , the spin peltier effect @xcite and the spin hall magnetoresistance ( smr ) @xcite . in the smr , both she and the ishe act in a concerted manner to allow electrical detection of the fmi magnetization direction . the sse refers to the conversion of thermal excitations of the magnetic order parameter ( spin waves or magnons ) into a spin current pumped into nm and detected by the ishe .
the sse and the smr have been investigated up to now only for a limited number of insulating ferrimagnets ( garnets and spinels ) with collinear magnetizations and recently , in an antiferromagnetic insulator @xcite . however , magnetic insulators come in a large variety of magnetic order . especially fascinating
are non - collinear magnets with competing magnetic interactions ( spin frustration ) induced by competing next - nearest neighbor exchange .
alternatively , the spin - orbit coupling , such as the moriya - dzyaloshinskii interaction favor complex spiral configurations and skyrmion order .
while the coupling of non - collinear magnetizations with spin , charge and heat transport is currently one of the hottest subjects in magnetism , its role in the smr and sse appears to have not been studied yet .
non - collinear magnetism emerges when the second nearest neighbor magnetic interactions are of the same order as the first one , generating geometrical frustration that favors spin canting .
various spin - spiral orders , like proper screw , cycloidal , longitudinal - conical and transverse - conical spiral , have been observed .
the cycloidal and the transverse - conical spiral orders break the inversion symmetry and induce a spontaneous electrical polarization , making these spiral magnetic systems multiferroic @xcite . here , we focus on non - collinear and multiferroic cocr@xmath0o@xmath1 ( cco ) thin films , reporting to the best of our knowledge first experimental observation of the smr and the sse in the pt@xmath5cco system for a wide range of temperatures including the ferrimagnetic and the spin - spiral phases .
cco is one of the rare multiferroic materials with linear magnetoelectric coupling @xcite .
it has a normal spinel structure with three sublattices .
the co@xmath9 ions are located exclusively in the tetrahedral sites ( forming one sublattice ) while cr@xmath10 ions reside in the octahedral sites ( in two sublattices ) .
the cr@xmath10 ions form a pyrochlore lattice with magnetic geometrical frustration @xcite .
bulk cco exhibits long - range collinear ferrimagnetic order below @xmath11 @xcite , with an easy axis of magnetization along the [ 001]-direction , as illustrated in fig .
[ fig:1](a ) and ( c ) . at @xmath12 ,
the ferrimagnetic long - range order adopts an additional short - range spiral order as illustrated in fig .
[ fig:1](b ) and ( c ) , which is known as the spin - spiral phase transition .
the conical spiral has 48@xmath13 , 71@xmath13 and 28@xmath13 cone angles with the [ 001]-direction for the co- , cr i- and cr ii- sublattices , respectively . below @xmath6 ,
magnetic spin - spirals move the oxygen atoms off - center due to the inverse dzyaloshinskii - moriya interaction @xcite , which results in the appearance of a spontaneous electrical polarization . at @xmath14 ,
the spin - spiral becomes commensurate with the lattice by the spin - lattice coupling , which is known as the spin lock - in transition . and ( b ) transverse conical spiral state for @xmath15 with an electrical polarization p. here , @xmath16 and @xmath6 are the curie transition temperature to a collinear ferrimagnetic phase and that the spin - spiral transition temperature , respectively .
( c ) the temperature dependence of the zero - field cooled magnetization of the cco target powder used for the film deposition for different applied magnetic fields .
@xmath7 is the spin lock - in magnetic transition temperature .
( d ) device configuration for the transverse resistance ( planar hall effect ) measurement of the pt film on top of cco .
, scaledwidth=70.0% ] here , we report a systematic study of the smr and the sse in the pt@xmath5cco bilayers from low to room temperature ( @xmath17 ) . at each temperature , we record the dependence on the angle of an in - plane applied magnetic field .
we observe strong effects of the cco magnetic order with largest signals in the spin lock - in phase at @xmath18 .
@xmath19-thick cco films were grown on ( 001 ) mgo substrates by pulsed laser deposition ( pld ) .
sintered ceramic cco targets @xcite were ablated with a krf excimer laser light with a wavelength of @xmath20 and a repetition frequency of @xmath21 . during the deposition , the mgo substrates are kept at @xmath22 in oxygen plasma atmosphere , having a base pressure of @xmath23 .
afterwards , the films were cooled down , with @xmath24 per minute in 0.5 bar o@xmath0 atmosphere . before further device fabrication , the films were annealed at @xmath25 for @xmath26 in an o@xmath0 atmosphere .
the crystal structure of the cco films was determined by x - ray diffraction ( xrd ) , where the rocking curves show high crystalline quality with a full width at half maximum ( fwhm ) below @xmath27 .
the magnetization of the cco films was measured by a squid magnetometer .
the films show an in - plane magnetic anisotropy with a coercive field @xmath28 around 2 t . for temperatures below @xmath29 , the magnetic transitions are hard to detect due to a large paramagnetic substrate background compared to the small magnetic moment of the thin film ( see appendix [ s3 ] ) .
nevertheless , squid measurements @xcite and x - ray resonance magnetic scattering both indicate the same magnetic transitions in thin films as reported for bulk cco . in order to demonstrate the magnetic transitions in cco , we measured the temperature dependence of the magnetization of bulk cco targets used for the film deposition .
[ fig:1](c ) shows a magnetization of the cco targets with the same transitions as reported in literature @xcite .
we carried out transport experiments on hall - bar structures patterned by electron beam lithography onto which a @xmath30-thick pt film was deposited by dc - sputtering , as shown in fig .
[ fig:1](d ) .
smr and sse have been measured simultaneously by lock - in detection @xcite .
using two stanford sr-830 lock - in amplifiers , the first and second harmonic voltage response were recorded separately . to minimize the background voltage
, we used the transverse instead of the longitudinal configuration , as schematically shown in fig.[fig:1](d ) , which is also referred to as planar hall voltage .
the smr signals scale linearly with the applied current and therefore detected in the pt resistance of the first harmonic response @xcite .
the current - induced sse scales quadratically with the applied current and therefore detected in the second harmonic response .
the angular dependence of the smr and the sse were studied by rotating an external magnetic field in the plane of the film @xcite that above @xmath16 induces a magnetization or ( if large enough ) aligns the direction of the cco magnetization below @xmath16 .
the in - plane angle @xmath31 of the magnetic field is defined relative to the current direction along the @xmath32-axis as indicated in fig .
[ fig:1](d ) .
all experiments were carried out in a quantum design ppms system , at magnetic fields @xmath33 and for temperatures from @xmath34to @xmath35 .
owing to the she , an ac - current through the pt hall - bar creates an ac spin accumulation at the pt@xmath5cco interface that can be partially absorbed or fully reflected , depending on the interface magnetization @xmath36 of the fmi ( see fig . [ fig:2](a ) and ( b ) ) .
the reflected spin currents generate an extra charge current via the ishe , thereby reducing the resistance . while the longitudinal resistance establishes the maximum modulation between these two configurations with @xmath37 or @xmath38 , the planar hall effect vanishes .
in contrast , when @xmath39 , the ishe is maximal @xcite , as sketched in fig . [ fig:2](c ) .
the additional emf scales with the applied current and is therefore , detected by the first harmonic transverse resistance @xmath40 @xcite , defined as @xmath41 , where @xmath42 is first harmonic signal of the lock - in amplifier generated by the applied ac current amplitude @xmath43 ( see appendix [ s1 ] for more details ) .
cco bilayers .
a charge current @xmath43 induces a spin current and thereby spin accumulation at the pt@xmath5cco interface by virtue of the she .
( a ) this spin accumulation is absorbed as a spin transfer torque when the magnetization @xmath36 is perpendicular to the current - induced spin polarization in pt , .
( b ) when @xmath36 is parallel to the spin accumulation , the spin current is reflected back into pt , where it generates an additional charge current , @xmath44 by the ishe .
( c ) when @xmath36 is at an angle @xmath31 to @xmath45 , the component of the spin accumulation perpendicular to @xmath36 is absorbed and the component parallel are reflected , leading to an extra charge current component normal to @xmath45 that is detected in the hall configuration ( d ) the angular dependence of the 1@xmath46 harmonic response in the transverse configuration , @xmath47 , for @xmath48ma at 5 k in an applied magnetic field of 6 t. the @xmath49 and @xmath50 curves illustrate the additive contributions from the ordinary hall effect and the smr .
( e ) , ( f ) and ( g ) show the temperature dependence of the smr , @xmath51 for @xmath52ma at 6 t , 4 t and 2 t , respectively . here , @xmath53 is the amplitude of the @xmath50 component from the smr .
@xmath16 , @xmath6 and @xmath7 are the ferrimagnetic , spin - spiral and spin lock - in magnetic transition temperatures , respectively .
the inset in ( e ) is a zoomed - in image of the smr signal below @xmath6 and @xmath7 at 6 t. , scaledwidth=80.0% ] the @xmath31 dependence of the first harmonic response of @xmath40 at 5 k in fig .
[ fig:2](d ) shows on top of the expected smr an ordinary hall effect ( ohe ) generated by a magnetic field component normal to the film due to a slight misalignment of the sample by an angle @xmath54 .
the ohe voltage has a @xmath55 angular dependence , where the phase @xmath56 is governed by the sample tilt direction .
a prefactor of @xmath57 at a current of @xmath58ma in a magnetic filed 6 t corresponds to a tilt of @xmath59 the ordinary hall voltage of pt@xmath5cco is nearly temperature independent and scales linearly with the applied current and magnetic field , as expected . after subtracting the ohe from @xmath40 , the anticipated @xmath50 dependence associated to
the smr remains @xcite , as illustrated in fig .
[ fig:2](d ) . the smr signal , @xmath51 ,
is defined as the amplitude of the @xmath50 component and plotted in fig .
[ fig:2](e ) , ( f ) and ( g ) as a function of temperature in magnetic fields of 6 t , 4 t and 2 t , respectively . exchanging the current and voltage probes in the hall bar in filed of 6
t leads to identical smr profiles , confirming that the interface magnetization is parallel to the applied field .
the smr of cco and yttrium iron garnet ( yig ) are compared in appendix [ s2 ] . below the collinear ferrimagnetic transition temperature @xmath4 , the smr signal increases with decreasing temperature .
the inset of fig .
[ fig:2](e ) recorded at a magnetic field of 6 t shows a distinct change of slope at the spin - spiral transition temperature to a non - collinear magnetic phase , @xmath2 . at @xmath15 ,
the smr signal is more than one order of magnitude larger than the signal observed at @xmath16 .
a further decrease in temperature below the spin lock - in transition temperature , @xmath3 , doubles the smr compared to @xmath6 .
this observation indicates that the exchange interaction between metal and magnet in the pt@xmath5cco system is more efficient in the non - collinear spiral phase than in the collinear ferrimagnetic phase .
the maximum smr signal is observed in the spin lock - in phase , when the period of the spin spiral becomes commensurate with the lattice . below @xmath16 , @xmath60
( see fig .
[ fig:3](a ) ) gives an excellent fit , where @xmath61 scales linearly with the applied magnetic field , as shown in fig .
[ fig:3](b ) .
@xmath61 should proportional to the interface spin - mixing conductance that vanishes with interface magnetization .
above @xmath62 all magnetization is generated by the applied magnetic field and @xmath61 is a measure of the interface paramagnetic susceptibility . for a magnetic interface ,
we anticipate a bilinear @xmath63 with a large slope at low magnetic fields that reflects the expulsion of magnetic domain walls .
the extrapolation of the high - field data should lead to a finite cut - off at zero magnetic fields ( as in studies of the anomalous hall effect ) .
however , the extrapolation of @xmath64 does not lead to a statistically significant @xmath65 . at present
we therefore can not confirm whether the observed smr signals reflect the paramagnetic susceptibility of a non - magnetic interface or a spontaneous interface magnetization texture . in order to shed more light on the @xmath66-dependence for @xmath67
, we compare the smr with the bulk magnetization , as shown in fig .
[ fig:1](c ) .
the @xmath16 , @xmath6 and @xmath7 transition temperatures established from the smr , closely corresponds to the transition temperatures observed in the bulk magnetization . in the spin - spiral and spin lock - in phases ,
the bulk magnetization does not depend much on temperature , as shown in fig .
[ fig:1](c ) , in stark contrast to the smr below @xmath6 ( fig .
[ fig:2](e ) ) .
the smr therefore does not directly reflect the cco bulk magnetization .
theoretically , the smr arises from the modulation of the spin current at the nm@xmath5fmi interface by the magnetization @xcite and is roughly proportional to the density of oriented magnet moments at the interface @xcite .
the smr enhancement below @xmath6 thus reflects an increased order of the magnetic moments at the interface with reduced temperature .
the moments at the interface can be either co ions , contributing to the ferrimagnetic component , or cr ions responsible for the cycloidal component of the spin - spiral .
the ferrimagnetic component of magnetization saturates around @xmath68 , therefore can not explain the enhanced smr below @xmath68 .
however , neutron scattering experiments show that the spiral component below @xmath2 strongly depends on temperature @xcite , similar to the smr signal .
we therefore venture that the development of the spiral order should be explained by a strongly temperature - dependent ordering of cr ions that can not be observed in the global magnetization .
this implies that by a simple transport experiment we can distinguish ferrimagnetic from cycloidal components of the interface magnetization . in a magnetic field of 6 t for @xmath69 .
the red curve shows the fit with curie law at low temperatures below t@xmath70 and the green curve shows the fit at the high temperature range above t@xmath70 with curie - weiss law ( for @xmath71 ) .
( b ) , ( c ) and ( d ) show the field dependence of fitting parameters @xmath61 , @xmath72 and @xmath73 , respectively . , scaledwidth=100.0% ] we observe a finite smr in the paramagnetic phase for @xmath8 complementing reports on spin pumping @xcite and the spin seebeck effect@xcite in the paramagnetic state . in cco the magnetic susceptibility , for @xmath8 , follows a curie - weiss law with a negative curie - weiss temperature @xmath74 which is evidence for antiferromagnetic correlations .
the high ratio @xmath75 indicates a significant magnetic frustration due to competing sublattice exchange interactions in cco @xcite , resulting in short range order above @xmath16 .
the smr signal above @xmath16 increases with temperature until it saturates to a constant value around room - temperature , as shown in fig .
[ fig:2](e - g ) .
the smr signal for @xmath8 provides evidence for an unusual interface magnetic susceptibility of our films .
we fitting the smr signals with the curie - weiss law , but a single fitting parameter @xmath72 does not capture the contributions from the molecular fields .
much unlike the bulk magnetic susceptibility , the smr signal is suppressed at @xmath16 which can be taking care of by introducing an additional parameter @xmath73 , as shown in fig .
[ fig:3](a ) . both fitting parameters @xmath72 and @xmath73 scale linearly with the applied magnetic field , as shown in fig .
[ fig:3](c ) and ( d ) .
this results support our conclusion drawn earlier that the smr signal can not be solely explained by the bulk magnetization of cco , even in the paramagnetic phase .
the temperature - independent background modelled by @xmath73 and the negative sign of @xmath72 remain unexplained by indicate that the short - range order reported for bulk cco above @xmath16 is importantly modified at the interface to pt .
we now discuss the sse signal observed in the second harmonic response .
it is caused by the joule heating of the pt hall - bar , generating a heat current into the ferromagnet which is absorbed by magnons .
this heat current is associated with a spin current polarized along the magnetization direction , which can be detected electrically by the ishe , as sketched in fig . [ fig:4](a ) . the second harmonic response
@xmath76 , where @xmath77 is the second harmonic signal of the lock - in amplifier at a phase set at @xmath78 @xcite .
the observed second harmonic response of the transverse resistance obeys the @xmath79 angular dependence anticipated for the sse , as shown in fig .
[ fig:4](b ) we define the amplitude of the @xmath80 for each temperature and magnetic field strength .
the sse signal for pt@xmath5cco has the same sign as for pt@xmath5yig ( see appendix [ s2 ] ) .
[ fig:4](c ) , ( d ) and ( e ) show the temperature dependence of @xmath81 in magnetic fields of 6 t , 4 t and 2 t , respectively . cco . a thermal gradient creates magnons at the interface by absorption of a spin current from pt with polarization @xmath82 , where @xmath36 is the cco interface magnetization@xmath83 the spin current generates an electromotive force @xmath84 by the ishe .
( b ) the angular dependence of the second harmonic response , @xmath85 at @xmath86 , for @xmath87 in an applied magnetic field of 6 t. ( c ) , ( d ) and ( e ) show the temperature dependence of @xmath81 at 6 t , 4 t and 2 t for @xmath87 , respectively .
the inset of ( c ) emphasizes the enhancement of the sse signal below the spin spiral ( @xmath6 ) and spin lock - in ( @xmath7 ) transition temperatures in a magnetic field of 6 t. ( f ) temperature dependence of @xmath81 in the non - linear current regime ( @xmath88 ) , with inset showing the magnetic field dependence of @xmath81 at @xmath89 .
, scaledwidth=70.0% ] in the collinear ferrimagnetic state at @xmath67 , the sse response increases with decreasing temperature .
[ fig:4](c ) gives evidence of a large sse enhancement below the spin spiral transition temperature @xmath6 and again below @xmath7 .
the sse signal below @xmath6 is five times larger than the signal observed at @xmath16 .
the inset of fig .
[ fig:4](c ) illustrates that the sse at @xmath90 increases by an order of magnitude from @xmath91 .
moreover below @xmath2 , the sse signal scales linearly with the applied magnetic field .
the sse is too noisy to provide as clear evidence for phase transitions at @xmath16 , @xmath6 and @xmath7 as the smr signals do .
conventional thermoelectric effects become small with decreasing temperatures , so what causes the observed remarkable enhancement at low temperatures ?
we can understand the temperature dependence of the sse signal by considering the contributions from different magnetic sublattices @xcite .
cco is a collinear ferrimagnet above @xmath6 , with three sublattices associated with co and oppositely polarized cr i and cr ii ion moments @xcite , but without any magnetization compensation below @xmath16 .
the magnetic sublattices contribute to the sse by correlated thermal fluctuations .
the coupled sublattices have acoustic ( ferromagnetic ) and optical ( antiferromagnetic ) modes .
the fundamental ferromagnetic modes govern the low energy excitations that are probed by ferromagnetic resonance and brillouin light scattering .
the optical modes are below @xmath16 shifted to higher energy by the exchange interaction . in the collinear ferrimagnetic state close to @xmath16 , the optical modes are still significantly occupied but the acoustic modes are slightly dominant .
the acoustic and optical modes contribute to the sse with different sign and therefore cancel to a large extent . by decreasing the temperature ,
the exchange splitting of the optical modes increases and therefore , become increasingly depleted .
the suppression of the thermal pumping of the optical modes therefore , leads to an apparent enhancement of the spin seebeck effect at lower temperature .
a similar mechanism explains the low temperature sign change of the sse of the ferrimagnetic insulator gd@xmath92fe@xmath93o@xmath93 ( gdig ) @xcite and can be used to understand the temperature dependence of the sse in yig at temperatures above 300 k @xcite .
in contrast to collinear magnetic order , the magnetization texture of a spin spiral is sensitively modulated by an external magnetic field @xcite .
this might contribute as well to the observed magnetic field dependence of the sse below @xmath6 , as shown in fig .
[ fig:4](c - e ) .
the corresponding increase in the sse with applied magnetic field below @xmath6 may persist until the co@xmath9 and cr@xmath10 momenta of @xmath94 are fully aligned .
no signs of magnetization saturation were observed at magnetic fields up to 30 t @xcite , so it would be very interesting to find out whether also the spin seebeck effect can be further enhanced in such high magnetic fields .
the observed correlation between the smr and sse provides evidence that the spin - mixing conductance , i.e. , the transport measure of the exchange interaction between ferromagnet and normal metal , does play an important role .
other factors , such as the magnetization damping and the magnon transport to the interface that is affected by the magnon - phonon interaction @xcite , may contribute to the observed enhancement in the sse below @xmath6 .
a quantitative description of the sse can be approached by atomistic spin simulations that take the full spin wave spectrum associated to the three sublattices with different cone angles , chirality , and damping parameters , as well the spin mixing conductances of the interface to pt into account .
above @xmath16 the sse can not be established at small heating current levels ( @xmath95 ) , as shown in fig .
[ fig:4](c ) .
however , at a larger current ( @xmath96 ) in a magnetic field of 4 t , a finite sse signal is detected above @xmath16 and up to room temperature , see fig .
[ fig:4](f ) . at these current levels , @xmath7 is shifted to lower temperature by @xmath97 due to sample heating , and non - linearities kick in , i.e. , the sse voltage @xmath98 .
the sse still scales linearly with the applied magnetic field as shown for @xmath89 in the inset of fig .
[ fig:4](f ) .
the presence of the sse above @xmath16 can be explained by the large longitudinal magnetic susceptibility of cco films that is responsible for the smr response as well @xcite . in summary , by lock - in detection we simultaneously measured the smr and sse in pt@xmath5cco bilayers .
the temperature dependence of the smr and , though less so , the sse , exposes distinct anomalies at the magnetic phase transitions .
a remarkable enhancement of both smr and sse signals is observed at low temperatures ( @xmath15 ) .
the smr is more than one order of magnitude larger at @xmath90 as compared to the signals around @xmath16 .
we relate the observed enhancement of the smr below @xmath6 to contributions from the cycloidal spiral projected onto the spin accumulation at the pt@xmath5cco interface .
the sse signal also increases by a factor of two when lowering temperature below @xmath6 .
the temperature dependence of the sse does not simply reflect the bulk magnetization ; instead the intricate magnetization dynamics of coupled sublattices needs to be considered .
our results suggest that the magnons from the complicated cr - sublattice magnetization texture plays a essential role in the sse .
vice versa , we establish that the smr and sse are powerful instruments that complement ferromagnetic resonance and neutron scattering techniques to analyze the magnetization dynamics of complex oxides including multiferroics .
we would like to acknowledge j. baas , h. bonder , m. de roosz and j. g. holstein for technical assistance .
this work is supported by the foundation for fundamental research on matter ( fom ) , nanonextnl , a micro- and nanotechnology consortium of the government of the netherlands and 130 partners , by nanolab nl , inspin eu - fp7-ict grant no 612759 and the zernike institute for advanced materials national research combination , grants - in - aid for scientific research ( grant nos .
25247056 , 25220910 , 26103006 ' '' '' , and dfg priority programme 1538 ( ba 2954/2 ) .
10 c. w. sandweg , y. kajiwara , a. v. chumak , a. a. serga , v. i. vasyuchka , m. b. jungfleisch , e. saitoh , and b. hillebrands .
spin pumping by parametrically excited exchange magnons .
, 106:216601 , may 2011 .
v. castel , n. vlietstra , b. j. van wees , and j. ben youssef . frequency and power dependence of spin - current emission by spin pumping in a thin - film @xmath99 system .
, 86:134419 , oct 2012 . c. hahn , g. de loubens , m. viret , o. klein , v. v. naletov , and j. ben youssef .
detection of microwave spin pumping using the inverse spin hall effect .
, 111:217204 , nov 2013 .
k. harii , t. an , y. kajiwara , k. ando , h. nakayama , t. yoshino , and e. saitoh .
frequency dependence of spin pumping in @xmath100 film .
, 109(11): , 2011 .
k. uchida , j. xiao , h. adachi , j. ohe , s. takahashi , j. ieda , t. ota , y. kajiwara , h. umezawa , h. kawai , g. e. w. bauer , s. maekawa , and e. saitoh .
spin seebeck insulator .
, 9(11):894897 , november 2010 . m. schreier , a. kamra , m. weiler , j. xiao , g. e. w. bauer , r. gross , and s. t. b. goennenwein .
magnon , phonon , and electron temperature profiles and the spin seebeck effect in magnetic insulator / normal metal hybrid structures . , 88:094410 , sep 2013 .
a. aqeel , i. j. vera - marun , b. j. van wees , and t. t. m. palstra .
surface sensitivity of the spin seebeck effect .
, 116(15):153705 , 2014 .
s. m. wu , j. e. pearson , and a. bhattacharya .
paramagnetic spin seebeck effect .
, 114:186602 , may 2015 .
j. flipse , f. k. dejene , d. wagenaar , g. e. w. bauer , j. ben youssef , and b. j. van wees .
observation of the spin peltier effect for magnetic insulators .
, 113:027601 , jul 2014 .
m. althammer , s. meyer , h. nakayama , m. schreier , s. altmannshofer , m. weiler , h. huebl , s. geprgs , m. opel , r. gross , d. meier , c. klewe , t. kuschel , j. schmalhorst , g. reiss , l. shen , a. gupta , y .- t .
chen , g. e. w. bauer , e. saitoh , and s. t. b. goennenwein .
quantitative study of the spin hall magnetoresistance in ferromagnetic insulator / normal metal hybrids .
, 87:224401 , jun 2013 .
n. vlietstra , j. shan , v. castel , b. j. van wees , and j. ben youssef .
spin - hall magnetoresistance in platinum on yttrium iron garnet : dependence on platinum thickness and in - plane / out - of - plane magnetization . , 87:184421 , may 2013 .
h. nakayama , m. althammer , y .- t .
chen , k. uchida , y. kajiwara , d. kikuchi , t. ohtani , s. geprgs , m. opel , s. takahashi , r. gross , g. e. w. bauer , s. t. b. goennenwein , and e. saitoh .
spin hall magnetoresistance induced by a nonequilibrium proximity effect . , 110:206601 , may 2013 .
n. vlietstra , j. shan , v. castel , j. ben youssef , g. e. w. bauer , and b. j. van wees .
exchange magnetic field torques in @xmath99 bilayers observed by the spin - hall magnetoresistance . , 103(3): , 2013 . c. hahn , g. de loubens , o. klein , m. viret , v. v. naletov , and j. ben youssef .
comparative measurements of inverse spin hall effects and magnetoresistance in @xmath99 and @xmath101 . , 87:174417 , may 2013 .
m. isasa , a. bedoya - pinto , s. vlez , f. golmar , f. snchez , l. e. hueso , j. fontcuberta , and f. casanova .
spin hall magnetoresistance at @xmath102 interfaces and texture effects .
, 105(14): , 2014 .
j. h. han , c. song , f. li , y. y. wang , g. y. wang , q. h. yang , and f. pan .
antiferromagnet - controlled spin current transport in @xmath103 hybrids .
, 90:144431 , oct 2014 .
m. fiebig .
revival of the magnetoelectric effect .
, 38(8):r123 , 2005 .
sang - wook cheong and maxim mostovoy .
multiferroics : a magnetic twist for ferroelectricity . , 6(1):1320 , 2007 .
y. tokura and s. seki .
multiferroics with spiral spin orders . , 22(14):15541565 , 2010 . r. ramesh and n. a. spaldin .
multiferroics : progress and prospects in thin films . , 6(1):219 , 2007 .
t. kimura and y. tokura .
magnetoelectric phase control in a magnetic system showing cycloidal / conical spin order .
, 20(43):434204 , 2008 .
y. yamasaki , s. miyasaka , y. kaneko , j .- p .
he , t. arima , and y. tokura .
magnetic reversal of the ferroelectric polarization in a multiferroic spinel oxide .
, 96:207204 , may 2006 .
y. j. choi , j. okamoto , d. j. huang , k. s. chao , h. j. lin , c. t. chen , m. van veenendaal , t. a. kaplan , and s - w .
thermally or magnetically induced polarization reversal in the multiferroic @xmath104 . , 102:067601 , feb 2009 .
k. tomiyasu , j. fukunaga , and h. suzuki .
magnetic short - range order and reentrant - spin - glass - like behavior in @xmath105 and @xmath106 by means of neutron scattering and magnetization measurements .
, 70:214434 , dec 2004 .
s. bordcs , d. varjas , i. kzsmrki , g. mihly , l. baldassarre , a. abouelsayed , c. a. kuntscher , k. ohgushi , and y. tokura .
magnetic - order - induced crystal symmetry lowering in @xmath107 ferrimagnetic spinels .
, 103:077205 , aug 2009 .
s. yang , h. x. bao , d. z. xue , c. zhou , j. h. gao , y. wang , j. q. wang , x. p. song , z. b. sun , x. b. ren , and k. otsuka .
magnetodielectric effect from the onset of ferrimagnetic transition in @xmath105 .
, 45(26):265001 , 2012 . g. lawes , b. melot , k. page , c. ederer , m. a. hayward , th .
proffen , and r. seshadri .
dielectric anomalies and spiral magnetic order in @xmath105 . , 74:024413 , jul 2006 .
k. singh , a. maignan , c. simon , and c. martin .
@xmath108 and @xmath105 spinels : multiferroicity in the collinear magnetic state ? , 99(17): , 2011 .
mufti n. , a. a. nugroho , g. r. blake , and t. t. m. palstra .
magnetodielectric coupling in frustrated spin systems : the spinels @xmath109 @xmath110 . ,
22(7 ) , 2010 - 2 - 24 .
k. r. choi , s. j. moon , t. kouh , i. b. shim , s. j. kim , and c. s. kim .
characterization of @xmath111 on @xmath112 grown by using pulsed laser deposition .
, 45(6):26102612 , june 2009 .
xiaoran liu , m. kareev , yanwei cao , jian liu , s. middey , d. meyers , j. w. freeland , and j. chakhalian .
electronic and magnetic properties of ( 1 1 1)-oriented @xmath105 epitaxial thin film . , 105(4): , 2014 .
v. tsurkan , s. zherlitsyn , s. yasin , v. felea , y. skourski , j. deisenhofer , h .- a .
krug von nidda , j. wosnitza , and a. loidl .
unconventional magnetostructural transition in @xmath105 at high magnetic fields .
, 110:115502 , mar 2013 .
n. vlietstra , j. shan , b. j. van wees , m. isasa , f. casanova , and j. ben youssef .
simultaneous detection of the spin - hall magnetoresistance and the spin - seebeck effect in platinum and tantalum on yttrium iron garnet .
, 90:174436 , nov 2014 .
m. schreier , n. roschewsky , e. dobler , s. meyer , h. huebl , r. gross , and s. t. b. goennenwein .
current heating induced spin seebeck effect .
, 103(24): , 2013 .
chen , s. takahashi , h. nakayama , m. althammer , s. t. b. goennenwein , e. saitoh , and g. e. w. bauer .
theory of spin hall magnetoresistance .
, 87:144411 , apr 2013 .
xingtao jia , kai liu , ke xia , and gerrit e. w. bauer .
spin transfer torque on magnetic insulators .
, 96(1):17005 , october 2011 .
y. shiomi and e. saitoh .
paramagnetic spin pumping .
, 113:266602 , dec 2014 .
y. ohnuma , h. adachi , e. saitoh , and s. maekawa .
spin seebeck effect in antiferromagnets and compensated ferrimagnets . , 87:014423 , jan 2013 .
s. geprgs , a. kehlberger , t. schulz , c. mix , f. d. coletta , s. meyer , a. kamra , m. althammer , g. jakob , h. huebl , r. gross , s. t.b .
goennenwein , and m. klui .
origin of the spin seebeck effect probed by temperature dependent measurements in @xmath113 . , 2014 .
k. uchida , t. kikkawa , a. miura , j. shiomi , and e. saitoh .
quantitative temperature dependence of longitudinal spin seebeck effect at high temperatures .
, 4:041023 , nov 2014 .
d. kamenskyi , h. engelkamp , t. fischer , m. uhlarz , j. wosnitza , b. p. gorshunov , g. a. komandin , a. s. prokhorov , m. dressel , a. a. bush , v. i. torgashev , and a. v. pronin .
observation of an intersublattice exchange magnon in @xmath111 and analysis of magnetic ordering .
, 87:134423 , apr 2013 .
v. i. torgashev , a. s. prokhorov , g. a. komandin , e. s. zhukova , v. b. anzin , v. m. talanov , l. m. rabkin , a. a. bush , m. dressel , and b. p. gorshunov .
magnetic and dielectric response of cobalt - chromium spinel @xmath105 in the terahertz frequency range .
, 54(2):350359 , 2012 .
s. r. marmion , m. ali , m. mclaren , d. a. williams , and b. j. hickey .
temperature dependence of spin hall magnetoresistance in thin @xmath99 films .
, 89:220404 , jun 2014 .
s. meyer , m. althammer , s. geprgs , m. opel , r. gross , and s. t. b. goennenwein .
temperature dependent spin transport properties of platinum inferred from spin hall magnetoresistance measurements .
, 104(24): , 2014 .
the temperature - dependent magnetization of the cco film in fig .
[ fig:6 ] as measured by squid magnetometry is evidence for a phase transition at @xmath4 to collinear ferrimagnet order , but the spin - spiral and spin lock - in transitions are not visible , because of the small magnetization of a thin film as explained in the main text .
the apparent increase in the magnetization below 20 k is probably caused by paramagnetic impurities in the substrate .
all measurements shown in the main text are carried out using a lock - in detection technique @xcite with @xmath114 ac current bias in the pt film .
the generated voltage can be expanded as @xmath115 where @xmath116 represent the @xmath117-th harmonic response . for @xmath118 , with angular frequency @xmath119 and rms value @xmath120 , the harmonic response coefficients @xmath116 are obtained by measuring the different frequency components @xmath121 by a lock - in amplifier .
the detected @xmath117-th harmonic response at a set phase @xmath56 can be written as @xmath122 focussing on first and second order response , we define the output voltage of the lock - in amplifier for the first and second harmonic responses by using eq .
( [ eq1 ] ) and eq .
( [ eq2 ] ) as : @xmath123{c}v_{1}=i_{0}r_{1}\\ v_{2}=i_{0}^{2}r_{2}/\sqrt{2}\end{array } \text { for } \begin{array } [ c]{c}\phi=0^{\text{o}}\\ \phi=-90^{\text{o}}\end{array } \label{eq34}\ ] ] the smr and the sse signals appear in the first and second harmonic responses in eq .
( [ eq34 ] ) , respectively . in the linear response regime at currents
@xmath124 , the smr scales linearly and the sse scales quadratically with the applied current as shown in fig .
[ fig:5](a , b ) .
however , at @xmath125 , the smr ( sse ) do no longer depend linearly ( quadratically ) on the current @xmath43 .
the results in the main text for the smr are gathered in the linear regime . at @xmath126
the sse signal decreases rapidly and we record the sse also at the high current @xmath127 , as shown in fig . [
fig:5](c ) . on @xmath128 due to the sse , generated by
current induced heating and ( b ) dependence of first harmonic contribution v@xmath129 , due to the smr , by an ac - current @xmath43 sent through the pt hall - bar in a magnetic field of 4 t. the smr scales linear and the sse scales quadratic i in the linear - regime ( @xmath130 ) .
( c ) the angular dependence of second harmonic response , @xmath131 , for @xmath132 , @xmath133 t at @xmath134 and @xmath135 .
, scaledwidth=80.0% ]
[ s2 ] to compare the smr and the sse response in pt@xmath5cco system to that of pt@xmath5yig , the angular dependence of pt@xmath5yig system is also systematically studied at different temperatures .
a @xmath30-thick pt hall - bar was deposited on a @xmath136 yig film by dc sputtering @xcite .
the pt hall - bar has a length of @xmath137 and a width of @xmath138 , with side contacts of @xmath139 width .
a @xmath140 thick yig , single - crystal , film is used and the film is grown by liquid phase epitaxy on a ( 111 ) gd@xmath92ga@xmath93o@xmath141 ( ggg ) substrate .
we observe the same sign of the smr signals for pt@xmath5cco and pt@xmath5yig , as expected .
a change of the sign of sse has been observed in compensated ferrimagnet @xcite but for both pt@xmath5yig and pt@xmath5cco system the sign remains the normal for all temperatures ( @xmath142 ) .
the temperature dependence of the smr and sse is observed to be very different in the pt@xmath5cco as compared to the pt@xmath5yig system . in pt@xmath5cco ,
both smr and the sse signals are enhanced at temperatures much lower than @xmath16 , as shown in fig .
[ fig:2](e ) and fig .
[ fig:4](c ) with maximal values around @xmath143 in the spin lock - in phase . in contrast ,
pt@xmath5yig displays conventional behavior , with both smr and sse larger at room temperature than at low temperatures . fig .
[ fig:7](a ) shows the temperature dependence pt@xmath5yig smr at relatively large current levels of @xmath144 .
the smr signal slightly increases when decreasing temperature until 150 k and decreases again when cooling even more @xcite .
the decrease in the smr signal in pt@xmath5yig has been ascribed to a decrease in the pt spin - hall angle with decreasing temperature @xcite .
the smr signal observed at 5 k is twice smaller than that at room - temperature .
the sse in pt@xmath5yig does not change much until @xmath145 , while a further decrease in temperature suppresses the sse as shown in fig .
[ fig:7](b ) .
a small increase in the sse is observed for @xmath146 .
the ordinary hall effect as observed in the first harmonic response of transverse resistance ( as shown in fig . [ fig:2](d ) ) is almost temperature independent in pt@xmath5cco system and scales linearly with the applied field as expected .
the ordinary hall effect is not observed in the pt@xmath5yig system because a much weaker field suffice to saturate the yig magnetization ( @xmath148 ) .
moreover , the smr signal observed in the pt@xmath5cco ( shown in fig . [ fig:2](e ) ) is smaller than that in the pt@xmath5yig system , resulting in a relatively larger contribution of the ordinary hall effect .
the smr response observed at 5 k in the pt@xmath5cco is more than one order of magnitude smaller than the signal observed in the pt@xmath5yig . | we report on the spin - hall magnetoresistance ( smr ) and spin seebeck effect ( sse ) in multiferroic cocr@xmath0o@xmath1 ( cco ) spinel thin films with pt contacts .
we observe a large enhancement of both signals below the spin - spiral ( @xmath2 ) and the spin lock - in transitions ( @xmath3 ) .
the smr and sse response in the spin lock - in phase are one order of magnitude larger than those observed at the ferrimagnetic transition temperature ( @xmath4 ) , which indicates that the interaction between spins at the pt@xmath5cco interface is more efficient in the non - collinear magnetic state below @xmath6 and @xmath7 . at @xmath8 ,
magnetic field - induced smr and sse signals are observed , which can be explained by a high interface susceptibility .
our results show that the spin transport at the pt@xmath5cco interface is sensitive to the magnetic phases but can not be explained solely by the bulk magnetization . |
since 1998 reliable cosmological data have been accumulated leading to the conclusion that our universe has entered recently into an accelerated expansion epoch .
the luminosity - distance data of supernovae ia ( sne ia ) has provided the main evidence .
hitherto explanations to this phenomenon are not satisfactory .
most models consider accelerated expansion as due to a component of the universe that behaves opposite to gravity , the so called _ dark energy _ ( de ) .
another unknown component of the universe is the _ dark matter _ ( dm ) , the missing mass necessary to held together galaxy clusters , also needed to explain the current large scale structure of the universe .
the preferred model to describe accelerated expansion is the lambda - cold - dark - matter ( @xmath7cdm ) .
however , cosmological constant has several theoretical drawbacks , like the discrepancies between the observed and theoretical values , i.e. the fine tuning problem ; the coincidence problem is another objection , meaning the fact that dark energy is nearly equal to the matter density just now . among many proposals to give a satisfactory answer
there are some models that consider de as a dynamical component , such as quintessence or k - essence ; in others proposals de appears as the result of the interaction of fundamental particles . in here
we adhere to the stream that considers dm and de as different manifestations of the same component , describing the dark sector as some kind of matter whose physical properties depend on the scale : behaving as dm at high densities and transforming into de at lower densities .
these unified dark matter models have the generalized chaplygin gas as its paradigmatic example @xcite , however these models present undesirable features like oscillations in the matter power spectrum and to mend this inconvenience it should be assumed the existence of nonadiabatic pressure perturbations . on the other side , in the context of inflation of the very early universe , it has been known since long time ago that an imperfect fluid with bulk viscosity can produce an acceleration in the expansion without falling back on a cosmological constant or some inflationary scalar field @xcite . inspired in those inflationary models
there are some recent @xcite and not so recent @xcite developments that assume a universe filled with a viscous single fluid .
the bulk viscosity contributes to the cosmic pressure and drives the accelerated expansion .
the origin of the bulk viscosity in a physical system is due to its deviations from the local thermodynamic equilibrium @xcite and different cooling rates @xcite . in a cosmological fluid ,
the bulk viscosity arises when the fluid expands ( or contracts ) too fast so that the system does not have enough time to restore the local thermodynamic equilibrium and then it appears an effective pressure restoring the system to its local thermodynamic equilibrium .
bulk viscosity can be seen as a measurement of this effective pressure .
when the fluid reaches again the thermal equilibrium then the bulk viscosity pressure ceases @xcite , @xcite .
so , in an accelerated expanding universe it may be natural to suppose that the expansion process is actually a collection of states out of thermal equilibrium in a small fraction of time that being modeled by a bulk viscous fluid might be a more realistic description of the accelerated universe today .
moreover , these models have in their favour that behave well when density perturbations come into play : the power spectrum for viscous matter is well behaved and consistent with large scale structure data @xcite .
in particular , it does not suffer from the oscillation problem of chaplygin models @xcite . on the other hand , since the earliest evidence of tight correlations in gamma - ray bursts spectral properties @xcite , the possibility arose of using grbs as standard candles .
being so grbs may open a window in redshift as far as @xmath8 @xcite , extending then the attainable range provided by sne ia observations . for an overview of most recent missions , see @xcite and references therein . at these epochs , @xmath8 ,
the universe was dominated by dark matter , from which follows that this tool is less sensitive to dark energy .
however for models where dark energy and matter are coupled @xcite or unified @xcite grbs might be a useful tool .
grbs were prevented of being used as standard candles because the intrinsic faintness of the nearby events , a fact that introduced a bias towards low redshifts of grb and therefore the extrapolation of their correlations to low - z events faced serious problems .
however this problem is solved if one uses sne ia data , in a combined calibration with grbs at low redshifts , allowing then grbs be considered as distance indicators @xcite . in particular , liang @xcite implemented a method to circumvent this objection , obtaining the distance modulus of grb at low redshift by interpolating from the hubble diagram of sne ia .
then grb relations are calibrated without assumming a particular cosmological model ; fair is to say that there is some criticism to the method @xcite . in this paper
we test the model ( presented in @xcite ) that considers a bulk viscous fluid as source of matter , with grbs data given in @xcite containing redshifts up to @xmath9 . in this way
we extend the previous sne ia probes of the model . in sec .
ii we introduced the main aspects of the bulk viscosity model ; in sec .
iii the calibration of grbs is addressed . in sec .
iv we apply the calibrated relation to higher redshifts and do the statistics to find the best fit for the bulk viscosity parameter .
we carry on the analysis with the observational data including sne ia ; bao for clusters with redshifts up approximately 2 , as well as cmb5-year and cmb7-year data from wmap ; this last probe tests the model up to the last scattering surface , about @xmath10 ; according to our analysis the model is not reliable to such far epoch , but we extend its applicability to epochs of formation of structure @xmath8 .
we discuss conclusions and perspectives in the last section .
in @xcite , the baryon and dark matter components are modeled by a pressureless fluid characterized by a constant bulk viscosity @xmath4 .
lying the bulk viscosity constant in the range @xmath5 , the model possesses many desirable properties , namely : holding the second law of thermodynamics , the derived age of the universe is in perfect agreement with the constraint of globular clusters , it has a big - bang followed by a decelerated expansion with a smooth transition to an accelerated expansion epoch in late times .
previous cosmological probes done by avelino and nucamendi include the sne ia gold 2006 sample ( 182 sne ia ) and bao in a quick fit @xcite .
the many good properties of the model justifies to complete with grbs the observational confrontation . in this paper
the model is tested for the first time using grbs , and updated tests using cmb five and seven years wmap data , as well as sne ia union2 sample and bao .
the results point out the strengths and what maybe be the limit of applicability of the model , as are the far far epochs of the universe , at very high redshifts of the recombination epoch . before going on
, we note that nucamendi - avelino @xcite addressed the possibility of extending the model of constant bulk viscosity to a model where the viscosity depends on @xmath11 as @xmath12 , where @xmath13 and @xmath14 are constants and @xmath11 is the hubble parameter . using the scp union2 sne ia data set composed of 557 type ia sne @xcite , they found that from all possible scenarios predicted by the model according to different values of the dimensionless bulk viscous coefficients @xmath15 and @xmath16 the preferred ones are two ( @xmath17 , @xmath18 ) with @xmath19 , and ( @xmath20 , @xmath21 ) . nevertheless , one of the disadvantages of the model parameterized by @xmath22 , is that the best estimated total bulk viscosity function @xmath23 is positive for redshifts @xmath24 and negative for @xmath25 latter implies a violation to the local second law of thermodynamics ( see @xcite for details ) .
then , the simplest model of one parameter @xmath26 turned out to be the best candidate to explain the present accelerated expansion .
for an imperfect fluid , the energy - momentum tensor with a first - order deviation from the thermodynamic equilibrium is : @xmath27 with @xmath28 where @xmath29 is the four - velocity vector of an observer who measures the effective pressure @xmath30 ; @xmath31 and @xmath32 are the pressure and density of the fluid respectively .
the term @xmath3 is the bulk viscous coefficient that arises in the fluid which is out of local thermodynamic equilibrium .
the conservation equation for the viscous fluid is @xmath33 in @xcite it has been considered a pressureless fluid ( @xmath34 ) , then @xmath35 .
we shall consider a spatially flat geometry for the friedmann - robertson - walker ( frw ) cosmology @xmath36 where the function @xmath37 is the scale factor .
the conservation equation [ eq : conservation ] now becomes , @xmath38 where @xmath39 is the hubble parameter , @xmath40 is the total matter density ( the dot means time derivative ) and @xmath41 . from the friedmann equations , the hubble parameter is given through @xmath42 where @xmath43 is the hubble constant today and it has been defined the dimensionless bulk viscous coefficient @xmath4 , the matter density parameter @xmath44 and the critical density today @xmath45 as @xmath46 so , the hubble parameter becomes @xmath47 ^ 2 .
\label{eq : hz}\ ] ] in this model the bulk viscous matter is the only component of the universe implying that the first friedmann equation for a flat universe , @xmath48 , evaluated today gives @xmath49 . with this ,
the hubble parameter becomes @xmath50 .
\label{eq : h}\ ] ] the function of the deceleration parameter @xmath51 is defined as @xmath52 with the second friedmann equation , @xmath53 , the term @xmath54 can be calculated . for a matter - dominated universe with bulk viscosity
, it is given by @xmath55 with the first friedmann equation , @xmath48 , we have @xmath56 then , substituting eq .
[ eq : rho ] and the definition of @xmath4 in [ eq : ddot : a ] we obtain @xmath57 and substituting [ eq : ddot : az ] into [ eq : q ] , we have @xmath58 with @xmath59 the eq .
[ eq : q(a ) ] is given as @xmath60 , \label{eq : q(a , zeta)}\ ] ] then the deceleration parameter today , is given by @xmath61 assuming the best estimated values for @xmath4 from table [ tabla:1 ] with grbs , the resulting deceleration parameter today is @xmath62 .
we shall analyze the model presented above . note that this is not the only proposal in this stream , but
many other similar models exist in the literature , based in similar ideas @xcite ; the conclusions we obtained then should be pertinent for models alike .
the main observables that can be measured when studying grbs are its spherical equivalent energy , its peak isotropic luminosity , the peak energy of its spectrum , the photon fluence , the energy fluence , the pulse duration and the redshift of its host galaxy . several empirical correlations among these variables can be established , see for example @xcite . the origin , reliability and dispersion of spectral energy correlations of grbs
have been debated these late years .
some of the pros and cons we comment next on the two most used correlations , @xmath63 ( amati ) relating the rest frame energy of the spectra @xmath64 and the isotropic energy emitted @xmath65 , as well as the @xmath66 ( ghirlanda ) , with @xmath67 that takes into account the non - isotropic release of energy of the grb .
butler _ et al . _
@xcite presented a very critical analysis on the observational selection effects , as the origin of the grbs energy correlations .
they concluded that rigorously , neither @xmath63 nor @xmath66 satisfy what should a physically - based correlation fulfil : to show a reduced scatter in the rest frame relative to the observer frame , and that must not persist if the assumed redshifts are scattered .
moreover , the discovery of some outliers to these relations has raised the suspicion that these correlations belong only to a sub - population of long grbs or that they are an artifact of the grbs detection process . in @xcite
it is pointed out that grbs correlations are strongly influenced by the number - biased against hard photons of the grbs detectors .
as far as the ghirlanda correlation @xmath66 , that in some analysis it turns out favored respect to the amati s , as the tighest of the grbs calibrations , it has the inconvenience that to be included in this relation , the grb afterglow must have an observed jet break in its light curve , thus , only a fraction of the observed events can contribute to establishing this relation .
amati @xcite contributes to robust the @xmath63 relation by testing two extremely energetic grbs ( grb080916c and grb090323 ) measured by _
fermi _ , showing that both events are fully consistent with @xmath63 correlation .
recently , in @xcite it is performed a rigorous analysis of the multivariate data , considering also selection effects of grbs , then the authors came to the conclusion that there exists a real , intrinsic correlation between @xmath63 , but not a narrow log - log and it is strongly detector dependent .
having in mind the previous warnings , we consider the @xmath63 relation along with data presented in @xcite and performing the pertinent calibration , we proceed to constrain the bulk viscosity model .
most calibrations take for granted a particular cosmological model , to remedy this circularity problem , liang @xcite proposed a cosmology - independent calibration method , consisting in calibrating at low redshifts using the sne ia data , mending so the few available low redshift grbs data ; the basic idea supporting the method is that light travels in the expanding universe in the same way no matter what its source be .
the typical spectrum of the prompt emission of grbs can be expressed as exponentially connected broken power - law , the so called band function @xcite
. then we can determine spectral peak energy @xmath68 , corresponding to the photon energy at maximum in @xmath69 spectra .
we shall apply the empirical relation @xmath70 that connects @xmath71 with the isotropic equivalent energy @xmath65 derived by amati @xcite , @xmath72 where @xmath73 is the bolometric fluence of gamma rays in the grb at redshift @xmath74 and @xmath75 is the luminosity distance of the grb ( of the hosting galaxy ) . using at low redshifts the 557 union2 sne ia data @xcite , wei ( 2010 ) derived the distance moduli for the 50 low - redshift ( @xmath76 ) grbs by using a cubic interpolation from the 557 union2 sne ia .
note that with the well - known relation @xmath77 one can convert the distance modulus @xmath78 into luminosity distance @xmath75 ( in units of mpc ) . from eq .
[ eq : eiso ] with the corresponding @xmath73 and luminosity distance @xmath75 , wei derived @xmath65 for these 50 grbs at @xmath76 .
furthermore , with the corresponding @xmath64 for these 50 grbs at @xmath76 , he found the best fit for the amati relation given as @xmath79 with @xmath80 wei s calibration was the result of an adjustment called bisector of the two ordinary least squares @xcite . instead of using wei s calibration we preferred to do our own fit with a minimum least square method , using the sample given in table 1 by wei @xcite , consisting of 50 low - redshift grbs ( @xmath76 ) .
we obtained the following best fit for the amati relation , @xmath81 the fit is shown in figure [ fig : a1 ] .
the errors are calculated using error propagation .
our calibration improves the one presented by capozziello and izzo @xcite , given by @xmath82 performed with the sne ia union sample of 307 objects .
moreover , being with errors of better than one part in @xmath83 , our calibration is also better than the amati correlation presented by liang @xcite , done using the 192 sne ia sample .
remind that at low redshifts the calibration is obtained using sne ia samples , and then extrapolated to the grbs with higher redshifts .
therefore the slight differences in calibrations are mainly the result of the different sne ia samples used at the low redshift calibration . extrapolating the calibrated amati relation to 59 high - redshift grbs ( @xmath84 ) given in table 2 by wei @xcite , with the corresponding @xmath68 and deriving @xmath65 from the calibrated amati relation , eq .
[ eq : newajusteeiso ] ( in what follows we named mb calibration ) , the distance moduli @xmath78 is obtained for the extended sample of 59 grbs at @xmath84 using eqs .
[ eq : eiso ] and [ eq : def_mu ] and the respective @xmath73 also reported in @xcite , in order to test the bulk viscosity model . to establish a comparison between results obtained using wei s calibration and mb calibration ,
we include both in tables [ tabla:1 ] , [ tabla:3 ] , [ tabla:4.1 ] , [ tabla:5 ] , [ tabla:6 ] .
otherwise , the grbs analysis was done with the mb calibration .
the 557 union2 sne ia data compiled in @xcite and the 59 calibrated grbs dataset in @xcite are given in terms of the distance modulus @xmath85 . the theoretical distance modulus is defined by @xmath86 on the other hand , given a parametrization @xmath87 depending on @xmath88 parameters @xmath89 , we can obtain the corresponding hubble free luminosity distance in a flat cosmology as @xmath90 using the maximum likelihood technique @xcite we can find the goodness of fit for the corresponding observed @xmath91 .
the goodness of fit corresponding to any set of parameters @xmath92 is determined by the probability distribution of @xmath92 , i.e. @xmath93 where @xmath94 ^ 2}{\sigma^2_{\mu_{obs}(z_i ) } } \label{eq : def_chi}\ ] ] and _ n _ is a normalization factor .
the parameters @xmath95 that minimize the @xmath96 expression [ eq : def_chi ] are the most probable parameter values ( the ` best fit ' ) and the corresponding @xmath97 gives an indication of the quality of the fitness for the given parametrization : the smaller @xmath98 is the better parametrization . for the case of the 557 union2 sne ia and the 59 calibrated grbs , @xmath96 is given for the model of bulk viscosity as @xmath99 ^ 2}{\sigma^2_{\mu_{obs}}(z_i)}. \label{eq : musnia}\ ] ] once constructed the @xmath100 function , eq .
[ eq : musnia ] , we minimize it to find the `` best fit '' for the free parameters of the model .
the probability distribution is rewritten as @xmath101 where n is a normalization factor . besides the sne ia and grbs , there are other observational data very relevant when testing cosmological models , so we consider the joint constraints from the latest observational data , combined with the sne ia and grbs , namely , the shift parameter @xmath102 from the wmap 5-year and 7-year data , and the distance parameter @xmath103 of the measurement of the baryon acoustic oscillation ( bao ) peak in the distribution of sdss luminous red galaxies . in the case of the anisotropies of the cosmic microwave background ( cmb ) ,
the observational parameter that is used for the construction of the @xmath104 is the shift parameter @xmath102 .
derived in the context of the bulk viscosity model @xmath102 is given by @xmath105 where the redshift of recombination is @xmath106 from wmap5 data @xcite and @xmath107 from wmap7 data @xcite .
@xmath96 for @xmath102-cmb is defined as @xmath108 ^ 2,\ ] ] where @xmath109 is the theoretical value predicted by the cosmological model and ( @xmath110 , @xmath111)= ( 1.710,0.019 ) from wmap5 @xcite and ( @xmath110 , @xmath111)= ( 1.725,0.018 ) from wmap7 @xcite . in the case of baryon acoustic oscillations ( bao )
, it is used the acoustic peak , given by @xmath112^{2/3},\ ] ] with @xmath113 , to construct the function @xmath96 as @xmath114 ^ 2,\ ] ] where and @xmath115 is the theoretical value predicted by the cosmological model . in @xcite , the value of @xmath103 has been determined to be @xmath116 . here
the scalar spectral index @xmath117 is taken to be 0.963 , which has been updated from wmap7 data @xcite .
so , the total @xmath118 is given by @xmath119 where @xmath100 is given in eq .
[ eq : musnia ] , @xmath120 and @xmath121 .
the best - fit model parameters are determined by minimizing the total @xmath122 . [
cols="^,^,^,^ " , ]
in the previous section , we have obtained the constraints of the bulk viscosity model with the latest observational data : the combination of 557 union2 sne ia data set in the range @xmath123 @xcite , 59 calibrated grbs data set at higher redshifts , the shift parameter @xmath102 from the wmap5 or wmap7 data @xcite , and the distance parameter @xmath103 of sdss luminous red galaxies @xcite .
as we mentioned above , the best - fit model parameters are determined by minimizing the total @xmath122 . for comparison , sne ia and sne ia + cmb + bao without grbs
have been used to show which is the contribution of grbs to the joint cosmological constraints .
in addition , some different data sets such as sne ia + bao and sne ia + bao + grbs have also been jointly considered to inspect if the bao or cmb data are responsible of spoiling the statistics . in figure [ fig:1 ] ,
we show the joint confidence regions in the ( @xmath4 , @xmath43 ) plane for the bulk viscosity model with @xmath124cte . using 557 sne ia + 59 grbs ( with high redshifts ) ,
the 1-@xmath125 confidence region for ( @xmath4 , @xmath43 ) of the bulk viscosity model is ( @xmath4 , @xmath43 ) = ( [email protected] , [email protected] ) with @xmath1270.9572 . for comparison ,
fitting results from the data without grbs are also given in figure [ fig:1 ] ( to the left ) .
with 557 sne ia , the best - fit values are ( @xmath4 , @xmath43 ) = ( [email protected] , [email protected] ) and @xmath1270.9830 .
we present in table [ tabla:1 ] the best - fit value of @xmath4 and @xmath43 with 1-@xmath125 uncertainties , @xmath98 and @xmath128 , as well as , the deceleration parameter @xmath6 and the redshift , @xmath129 , of the transition between the decelerated - accelerated expansion epochs , as functions of @xmath4 ; the deceleration parameter is according to the @xmath7cdm prediction .
moreover , from these values for @xmath4 and @xmath43 , we see that no significative difference arise from probing with sne ia or with sne ia + grbs .
then , the validity of the constant bulk viscosity model is extended to redshifts of @xmath130 . in table
[ tabla:2 ] , we present the best - fit value of @xmath4 and @xmath43 with 1-@xmath125 uncertainties , and @xmath98 , @xmath128 , as well as , @xmath6 and @xmath129 but this time using the combination sne ia + cmb5 + bao , sne ia + cmb7 + bao , sne ia + cmb5 + bao + grbs and sne ia + cmb7 + bao + grbs , respectively . in all the cases , we found @xmath131 that is a value unacceptable since it is associated to violation of the second law of thermodynamics , however the inclusion of grbs significantly improves the statistics . in table
[ tabla:2 ] , check first and third column and second and fourth , including grbs we obtained a better value in @xmath128 the negative values of @xmath4 has as consequence that @xmath132 ( see eq .
[ eq : q(a=1,zeta ) ] ) and @xmath133 becomes negative .
moreover , being @xmath131 the pressure turns out to be positive since @xmath134 .
the meaning of this is that we can not extrapolate the viscous fluid behaviour to so far epochs .
we can conclude that the model has limited applicability when extrapolated to very large redshifts like that of the recombination epoch , i.e. when the statistical tests include the cmb .
notice that all the parameter suffer a drastic change in their estimated values , the unreliability of them is measured by the increasing in @xmath128 and @xmath98 as 2 and as doubled data , respectively .
however , the statistics is better when included grbs than just sne ia , i.e. sne ia + cmb + bao is worst than sne ia + cmb + bao + grbs . in the following tables
[ tabla:3 ] , [ tabla:4 ] and [ tabla:4.1 ] are presented the best - fit values of @xmath4 with 1-@xmath125 uncertainties , and @xmath98 , @xmath128 , as well as @xmath129 and @xmath6 for the model of bulk viscosity assuming a constant prior over @xmath43 . in table
[ tabla:3 ] are shown the results obtained using sne ia and sne ia + grbs , which are very similar to those obtained without prior over @xmath43 .
the effect of assuming a constant prior on @xmath43 , with and without grbs , is a best value of @xmath135 in table [ tabla:4 ] the results are using sne ia + cmb5 + bao , sne ia + cmb7 + bao , sne ia + cmb5 + bao + grbs and sne ia + cmb7 + bao + grbs .
the obtained values are almost the same with or without assuming a constant prior on @xmath43 but the best - fit is obtained without prior , and statistics is better when grbs are included .
until now we have seen that when we consider only sne ia + grbs we obtain @xmath136 and a reasonable value for the hubble constant @xmath43 .
however , when the same analysis is performed using sne ia , cmb5 or cmb7 , and bao with or without grbs , we obtain negative values of @xmath4 that disagree with the second law of thermodynamics and moreover the estimation for @xmath6 and @xmath129 is not good and also @xmath128 gets worst . since the redshift of cmb corresponds to earlier epochs of the universe while bao refers to more recent ones ( @xmath137 ) ,
our guess is that it is the inclusion of cmb that ruins our statistics .
the aim of table [ tabla:4.1 ] is to prove it .
we derived the constraints for the model of bulk viscosity using sne ia + bao and sne ia + bao + grbs and we found , in both cases , that the value of @xmath4 is in perfect agreement with the value predicted by the model to have an accelerated expansion epoch , that is to say , @xmath138 , and also with this value of @xmath4 we have an accelerated universe in the present with a transition between decelerated - accelerated expansion at @xmath139 .
this result agrees with the analysis on perturbation dynamics and the calculated matter power spectrum done by zimdahl @xcite on the bulk viscosity model , that turned out compatible with the data from 2dfgrs and sdss surveys .
finally , we analyze and constraint the viability of the model assuming a dirac delta prior over @xmath43 located at @xmath14070.5 km s@xmath141 mpc@xmath141 as reported by 5-year wmap data and at @xmath14073.8 km
s@xmath141 mpc@xmath141 as suggested by the observations of the hst , the results are shown from table [ tabla:5 ] to table [ tabla:8 ] .
again , the results obtained with @xmath14070.5 km s@xmath141 mpc@xmath141 using sne ia and using the joint sne ia + grbs , are in agreement with the model and good values are obtained for @xmath129 and @xmath6 .
however when we use @xmath14073.8 km s@xmath141 mpc@xmath141 we obtained values for @xmath4 in the expected range but with a slight increment in the value of @xmath128 this is shown in table [ tabla:5 ] and [ tabla:6 ] .
on the other hand , when we used sne ia , cmb data ( wmap5 or wmap7 ) and bao with or without grbs , assuming a dirac delta prior over @xmath43 located at @xmath14070.5
km s@xmath141 mpc@xmath141 or @xmath14073.8 km s@xmath141 mpc@xmath141 , we found values for @xmath4 in the range suggested by the model but with bad statistic , which is reflected in the value of @xmath128 moreover , since @xmath142 , then @xmath133 turns out to be enormous , and @xmath143 , for these reasons we conclude that the statistics for a delta dirac prior on @xmath43 is not reliable .
it should be noted that when we use sne ia + cmb5 + bao + grbs or sne ia + cmb7 + bao + grbs , i.e. when grbs are included , significantly improve the value of @xmath128 , but anyway , we have a worse fit to data each time that cmb is considered .
see tables [ tabla:7 ] and [ tabla:8 ] .
lastly , if we compare the results obtained using wei s calibration or mb calibration for grbs , as shown in tables [ tabla:1 ] , [ tabla:3 ] , [ tabla:4.1 ] , [ tabla:5 ] and [ tabla:6 ] , we notice that in most cases @xmath144 is closer to 1 when using mb calibration , that is to say , the latest calibration provide a better fit .
modeling the dark sector of the universe with a bulk viscous fluid with pressure @xmath145 , with a constant bulk viscosity @xmath146 , is tested with several cosmological data : sne ia , cmb and bao jointly with the calibrated sample of grbs .
the model is in good agreement with observational data from sn ia , bao and grbs , up to the range of @xmath9 with a chi - square statistics of @xmath147 . probing with sne ia and grbs we obtained as the present hubble parameter @xmath148 , the deceleration parameter @xmath149 , both in good agreement with the accepted values ;
as the redshift transition parameter we obtain @xmath150 .
when the baryon acoustic oscillations ( bao ) data are included in the data - model confrontation , we obtain a viscosity parameter @xmath151 , included in the correct range , and improved values for @xmath152 and @xmath147 .
the results with bao confirm the analysis performed by hiplito - ricaldi ( 2009 ) @xcite on the matter power spectrum showing the compatibility of the bulk viscosity models with the 2dfgrs and sdss data .
however the model does not pass the cmb test , exhibiting @xmath153 for cmb5 and @xmath154 for cmb7 , besides a negative bulk viscosity constant , violating so the second law of thermodynamics .
these results are consequence of neglecting a radiative component in the universe , that in so far epochs as the recombination , plays an important role .
thus , when testing so far epochs a radiation component should be added to the universe content , a component that is conserved independently of the rest of matter considered . to conclude , constant bulk viscosity matter reproduce observational sn ia , grbs and bao data in good agreement ( @xmath147 ) , being then a reliable model not only for the recent state of the universe , but for epochs where structure already existed ( @xmath9 ) .
it is an open questions if the model could account for the cmb data if a radiation component is included in the considered matter content of the universe .
o. bertolami , f. gil - pedro and m. le delliou , _ dark energy - dark matter interaction and putative violation of the equivalence principle from the abell cluster a586 _ , phys .
b * 654 * ( 2007 ) 165 [ arxiv : astro - ph/0703462 ] .
g. ghirlanda , g. ghisellini and d. lazzati , _ the collimation - corrected gamma - ray burst energies correlate with the peak energy of their @xmath69 spectrum _ , astrophys .
j. * 616 * ( 2004 ) 331 [ arxiv : astro - ph/0405602 ] .
o. bertolami and p. t. silva , _ gamma - ray bursts as dark energy - matter probes in the context of the generalized chaplygin gas model _ , mon . not .
* 365 * ( 2006 ) 1149 [ arxiv : astro - ph/0507192 ] . c. firmani , g. ghisellini , v. avila - reese and g. ghirlanda , _ discovery of a tight correlation among the prompt emission properties of long gamma - ray bursts _ , mon . not .
* 370 * ( 2006 ) 185 [ arxiv : astro - ph/0605073 ] .
a. shahmoradi and r. j. nemiroff , _ the possible impact of gamma - ray bursts detector thresholds on cosmological standard candles _ ,
mon . not .
* 411 * ( 2011 ) 1843 [ arxiv : astro - ph/0904.1464 ] . | in this paper we extend the range of consistency of a constant bulk viscosity model to redshifts up to @xmath0 . in this model
the dark sector of the cosmic substratum is a viscous fluid with pressure @xmath1 , where @xmath2 is the fluid - expansion scalar and @xmath3 is the coefficient of bulk viscosity . using the sample of 59 high - redshift grbs reported by wei ( 2010 )
, we calibrate grbs at low redshifts with the union 2 sample of sne ia , avoiding then the circularity problem .
testing the constant bulk viscosity model with grbs we found the best fit for the viscosity parameter @xmath4 in the range @xmath5 , being so consistent with previous probes ; we also determined the deceleration parameter @xmath6 and the redshift of transition to accelerated expansion .
besides we present an updated analysis of the model with cmb5-year data and cmb7-year data , as well as with the baryon acoustic peak bao . from the statistics with cmb it turns out that the model does not describe in a feasible way the far far epoch of recombination of the universe , but is in very good concordance for epochs as far as @xmath0 till present . |
most methods for estimating translation models from parallel texts ( bitexts ) start with the following intuition : words that are translations of each other are more likely to appear in corresponding bitext regions than other pairs of words .
the intuition is simple , but its correct exploitation turns out to be rather subtle .
most of the literature on translation model estimation presumes that corresponding regions of the input bitexts are represented by neatly aligned segments . as discovered by @xcite ,
most of the bitexts available today are not easy to align . moreover ,
imposing an alignment relation on such bitexts is inefficient , because alignments can not capture crossing correspondences among text segments .
@xcite proposed methods for producing general bitext maps for arbitrary bitexts .
the present report shows how to use bitext maps and other information to construct a model of co - occurrence .
a * model of co - occurrence * is a boolean predicate , which indicates whether a given pair of word _ tokens _ co - occur in corresponding regions of the bitext space .
co - occurrence is a precondition for the possibility that two tokens might be mutual translations .
models of co - occurrence are the glue that binds methods for mapping bitext correspondence with methods for estimating translation models into an integrated system for exploiting parallel texts . when the model of co - occurrence is modularized away from the translation model , it also becomes easier to study translation model estimation methods _ per se_. different models of co -
occurrence are possible , depending on the kind of bitext map that is available , the language - specific information that is available , and the assumptions made about the nature of translational equivalence .
the following three sections explore these three variables .
by definition of `` mutual translations , '' corresponding regions of a text and its translation will contain word token pairs that are mutual translations .
therefore , a general representation of bitext correspondence is the natural concept on which to build a model of where mutual translations co - occur .
the most general representation of bitext correspondence is a bitext map @xcite .
token pairs whose co - ordinates are part of the true bitext map ( tbm ) are mutual translations , by definition of the tbm .
the likelihood that two tokens are mutual translations is inversely correlated with the distance between the tokens co - ordinate in the bitext space and the interpolated tbm .
it may be possible to develop translation model estimation methods that take into account a probabilistic model of co - occurrence . however , all the models in the literature are based on a boolean co - occurrence model they want to know either that two tokens co - occur or that they do not
. a boolean co - occurrence predicate can be defined by setting a threshold @xmath0 on the distance from the interpolated bitext map .
any token pair whose co - ordinate is closer than @xmath0 to the bitext map would be considered to co - occur by this predicate .
the optimal value of @xmath0 varies with the language pair , the bitext genre and the application .
figure [ dcooc ] illustrates what i will call the * distance - based model of co - occurrence*. @xcite were the first to use a distance - based model of co - occurrence , although they measured the distance in words rather than in characters .
general bitext mapping algorithms are a recent invention .
so far , most researchers interested in co - occurrence of mutual translations have relied on bitexts where sentence boundaries ( or other text unit boundaries ) were easy to find @xcite .
aligned text segments suggest a * boundary - based model of co - occurrence * , illustrated in figure [ scooc ] . for bitexts involving languages with similar word order , a more accurate * combined model of co - occurrence *
can be built using both segment boundary information and the map - distance threshold .
as shown in figure [ bcooc ] , each of these constraints eliminates the noise from a characteristic region of the bitext space .
both the boundary - based and distance - based constraints restrict the region of the bitext space where tokens may be considered to co - occur . yet , these constraints do not answer the question of how to count co - occurrences within the restricted regions .
it is somewhat surprising that this is a question at all , and most authors ignore it .
however , when authors specify their algorithms in sufficient detail to answer this question , the most common answer ( given , , by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) turns out to be unsound .
the problem is easiest to illustrate under the boundary - based model of co - occurrence .
given two aligned text segments , the naive way to count co - occurrences is @xmath1 where @xmath2 and @xmath3 are the frequencies of occurrence of @xmath4 and @xmath5 in their respective segments . for
many * u * and * v * , @xmath2 and @xmath3 are either 0 or 1 , and equation [ naivecooc ] returns 1 just in case both words occur .
the problem arises when @xmath6 and @xmath7 .
for example , if @xmath8 , then according to equation [ naivecooc ] , @xmath9 ! if the two aligned segments are really translations of each other , then it is most likely that each of the occurrences of @xmath4 is a translation of just one of the occurrences of @xmath5 .
although it may not be known which of the 3 @xmath10 s each @xmath11 corresponds to , the number of times that @xmath12 and @xmath5 co - occur as possible translations of each other in that segment pair must be 3 .
there are various ways to arrive at @xmath13 .
two of the simplest ways are @xmath14\ ] ] and @xmath15 .\ ] ] equation [ mincount ] is based on the simplifying assumption that each word is translated to at most one other word . equation [ maxcount ] is based on the simplifying assumption that each word is translated to at least one other word . either simplifying assumption results in more plausible co - occurrence counts than the naive method in equation [ naivecooc ] . counting co - occurrences is more difficult under a distance - based co - occurrence model , because there are no aligned segments and consequently no useful definition for @xmath16 and @xmath17 . furthermore , under a distance - based co - occurrence model , the co - occurrence relation is not transitive . _
e.g. _ , it is possible that @xmath18 co - occurs with @xmath19 , @xmath19 co - occurs with @xmath20 , @xmath20 co - occurs with @xmath21 , but @xmath18 does not co - occur with @xmath21 .
the correct counting method becomes clearer if the problem is recast in graph - theoretic terms .
let the words in each half of the bitext represent the vertices on one side of a bipartite graph .
let there be edges between each pair of words whose co - ordinates are closer than @xmath0 to the bitext map .
now , under the `` at most one '' assumption of equation [ mincount ] , each co - occurrence is represented by an edge in the graph s maximum matching . under the `` at least one '' assumption of equation [ maxcount ] , each co - occurrence is represented by an edge in the graph s smallest vertex cover .
maximum matching can be computed in polynomial time for any graph @xcite .
vertex cover can be solved in polynomial time for bipartite graphs .
it is of no importance that maximum matchings and minimum vertex covers may be non - unique by definition , all solutions have the same number of edges , and this number is the correct co - occurrence count .
co - occurrence is a universal precondition for translational equivalence among word tokens in bitexts .
other preconditions may be imposed if certain language - specific resources are available @xcite .
for example , parts of speech tend to be preserved in translation @xcite . if part - of - speech taggers are available for both languages in a bitext , and if cases where one part of speech is translated to another are not important for the intended application , then we can rule out the possibility of translational equivalence for all token pairs involving different parts of speech
. a more obvious source of language - specific information is a machine - readable bilingual dictionary ( mrbd ) .
if token @xmath22 in one half of the bitext is found to co - occur with token @xmath23 in the other half , and @xmath24 is an entry in the mrbd , then it is highly likely that the tokens @xmath22 and @xmath23 are indeed mutual translations . in this case
, there is no point considering the co - occurrence of @xmath22 or @xmath23 with any other token .
similarly exclusive candidacy can be granted to cognate token pairs @xcite .
most published translation models treat co - occurrence counts as counts of potential link tokens @xcite .
more accurate models may result if the co - occurrence counts are biased with language - specific knowledge . without loss of generality , whenever translation models refer to co - occurrence counts , they can refer to co - occurrence counts that have been filtered using whatever language - specific resources happen to be available .
it does not matter if there are dependencies among the different knowledge sources , as long as each is used as a simple filter on the co - occurrence relation @xcite .
in this short report , i have investigated methods for modeling word token co - occurrence in parallel texts ( bitexts ) .
models of co - occurrence are a precursor to all the most accurate translation models in the literature .
so far , most researchers have relied on only a restricted form of co - occurrence , based on a restricted kind of bitext map , applicable to only a limited class of bitexts .
a more general co - occurrence model can be based on any bitext map , and thus on any bitext . the correct method for counting the number of times that two words co - occur turns out to be rather subtle , especially for more general co - occurrence models . as noted in section [ countmethod ] , many published translation models have been based on flawed models of co - occurrence .
this report has exposed the flaw and has shown how to fix it .
( 1995 ) `` a pattern matching method for finding noun and proper noun translations from noisy parallel corpora , '' _ proceedings of the 33rd annual meeting of the association for computational linguistics_. boston , ma .
a. kumano & h. hirakawa .
( 1994 ) `` building an mt dictionary from parallel texts based on linguistic and statistical information , '' _ proceedings of the 15th international conference on computational linguistics_. kyoto , japan .
j. kupiec .
( 1993 ) `` an algorithm for finding noun phrase correspondences in bilingual corpora , '' _ proceedings of the 31st annual meeting of the association for computational linguistics_. columbus , oh . h. papageorgiou , l. cranias & s. piperidis .
( 1994 ) `` automatic alignment in parallel corpora , '' _ proceedings of the 32nd annual meeting of the association for computational linguistics ( student session)_. las cruces , nm . | a * model of co - occurrence * in bitext is a boolean predicate that indicates whether a given pair of word _ tokens _ co - occur in corresponding regions of the bitext space .
co - occurrence is a precondition for the possibility that two tokens might be mutual translations
. models of co - occurrence are the glue that binds methods for mapping bitext correspondence with methods for estimating translation models into an integrated system for exploiting parallel texts
. different models of co - occurrence are possible , depending on the kind of bitext map that is available , the language - specific information that is available , and the assumptions made about the nature of translational equivalence .
although most statistical translation models are based on models of co - occurrence , modeling co - occurrence correctly is more difficult than it may at first appear . |
understanding the formation and subsequent evolution of the galactic bulge is important both for interpreting observations of extragalactic populations and for constraining galaxy chemodynamical formation models .
recent large sample spectroscopic surveys , such as the bulge radial velocity assay ( brava ; rich et al . 2007a ; howard et al . 2008 ; howard et al .
2009 ; kunder et al .
2012 ) , the abundances and radial velocity galactic origins survey ( argos ; freeman et al . 2013 ; ness et al . 2012 ; 2013b ) , the apache point observatory galactic evolution experiment ( apogee ; majewski et al . 2010 ; nidever et al .
2012 ) , and the giraffe inner bulge survey ( gibs ; zoccali et al .
2014 ) provide a coherent view of the bulge as a barred , triaxial system exhibiting cylindrical rotation .
photometric and star count studies have also discovered a double red clump along some bulge sight lines that traces out an x shaped structure ( mcwilliam & zoccali 2010 ; nataf et al . 2010 ; saito et al .
this structure appears to be dominated by stars with [ fe / h]@xmath40.5 on bar supporting orbits ( soto et al .
2007 ; babusiaux et al . 2010 ; ness et al . 2012 ; uttenthaler et al . 2012
; but see also nataf et al .
2014 ) .
inclusive with these data are detailed composition analyses of field stars from moderate and high resolution spectroscopy ( mcwilliam & rich 1994 ; ramrez et al . 2000 ; rich & origlia 2005 ; cunha & smith 2006 ; fulbright et al .
2006 ; zoccali et al .
2006 ; fulbright et al . 2007 ; lecureur et al . 2007 ;
rich et al . 2007b ; cunha et al . 2008 ; melndez et al . 2008 ; zoccali et al .
2008 ; alves brito et al . 2010 ; bensby et al .
2010a ; ryde et al .
2010 ; bensby et al . 2011 ; gonzalez et al . 2011 ; hill et al . 2011 ; johnson et al . 2011 ; johnson et al . 2012
; rich et al . 2012 ; uttenthaler et al . 2012 ; barbuy et al . 2013 ; bensby et al . 2013 ; garca prez et al . 2013 ; johnson et al . 2013a ; ness et al . 2013a ; jnsson et al . 2014 ) finding , at least in a general sense , that the bulge is composed of stars spanning more than a factor of 100 in [ fe / h]@xmath6log(n@xmath7/n@xmath8)@xmath9log(n@xmath7/n@xmath8)@xmath10 and log @xmath11(a)@xmath6log(n@xmath7/n@xmath12)+12.0 for elements a and b. ] , that bulge stars are uniformly enhanced in their [ @xmath3/fe ] ratios at low metallicity relative to the thin disk , and that the median [ fe / h ] along bulge sight lines decreases as a function of increasing galactic latitude ( i.e. , there is a metallicity gradient ) .
the enhanced [ @xmath3/fe ] abundances , coupled with the low [ la / eu ] ratios of bulge stars ( mcwilliam et al .
2010 ; johnson et al .
2012 ) , are consistent with the notion that the bulge formed rapidly ( @xmath1313 gyr ) .
in fact , the bulge appears uniformly old ( @xmath110 gyr ) in age studies based on color magnitude diagram analyses ( e.g. , ortolani et al .
1995 ; zoccali et al . 2003
; clarkson et al . 2008 ; valenti et al . 2013
; but see also ness et al .
2014 ) , and clarkson et al .
( 2011 ) estimate from the blue straggler population in an inner bulge field that a truly young ( @xmath145 gyr ) population should not constitute more than @xmath13.4@xmath15 of the bulge .
in contrast , ages derived from microlensed dwarf studies ( e.g. , bensby et al . 2013 ) find that while all metal poor bulge stars are uniformly old , @xmath1525@xmath15 of metal rich stars , at least near the galactic plane , may be only @xmath128 gyr in age .
while the observational data continue to grow , the difficult task of assembling the pieces into a fully self consistent model of the bulge s formation remains open .
the chemodynamical bulge data are challenging to interpret .
the bulge s predominantly old age , enhanced [ @xmath3/fe ] ratios , vertical metallicity gradient , and the existence of possible primordial building blocks " such as terzan 5 ( e.g. , ferraro et al 2009 ; origlia et al . 2011 ; 2013 )
are more consistent with the classical , merger built formation scenario .
however , the bulge s boxy x shape , similar composition characteristics to at least the thick disk , and cylindrical rotation profile suggest that the bulge formed via secular evolution from a buckling disk instability and may be a pseudobulge " ( e.g. , kormendy & kennicutt 2004 ; but see also zoccali et al .
2014 ) . while shen et al .
( 2010 ) rule out a classical bulge component that exceeds @xmath18@xmath15 of the disk mass , it may still be possible for a bar to form within a pre existing classical bulge ( e.g. , saha et al .
additionally , evidence such as the metallicity gradient may not be unique to the classical bulge scenario , and may be consistent with a secular evolution model in which a radial metallicity gradient in the buckling disk is transformed into a vertical gradient in the resultant bar ( martinez valpuesta & gerhard 2013 ) .
the bulge may also be composed of at least two stellar populations with different composition and kinematics ( babusiaux et al .
2010 ; hill et al .
2011 ; bensby et al . 2011
; bensby et al . 2013 ; ness et al .
however , at the moment the exact nature of these potentially distinct stellar populations is far from certain .
although most of the chemical abundance work mentioned previously has focused on the [ fe / h ] and [ @xmath3/fe ] ratios in comparison with the thin and thick disks , the light odd z and fe peak ( and also neutron capture ) elements also provide discriminatory power between models and other stellar populations ( e.g. , see kobayashi et al . 2011 ; their figure 14 ) .
peak elements in particular are useful as they may be sensitive to formation environment and metallicity .
for example , the metallicity dependent yields and increased contributions of massive stars are predicted to produce enhanced [ cu / fe ] and [ zn / fe ] ratios in the bulge compared to the local disk . similarly , if the bulge formed with a significantly flatter initial mass function ( imf ) than the disk then bulge stars should exhibit very large [ co / fe ] and [ zn / fe ] ratios ( nomoto et al .
therefore , here we measure abundances of the fe
peak elements cr , fe , co , ni , and cu , in addition to the light odd z and @xmath3elements o , na , mg , al , si , and ca , in 156 red giant branch ( rgb ) stars in two galactic bulge fields at ( l , b)=(@xmath05.25,3.02 ) and ( 0,12 ) , and compare the abundance ratios with other bulge fields , the galactic disk , and chemical enrichment models .
the flames giraffe spectra for this project are based on data obtained from the eso science archive facility under request number 51251 , which are based on observations collected at the european southern observatory , paranal , chile ( eso program 073.b0074 ) .
details regarding the selection of targets and input parameters ( e.g. , photometry and astrometry ) are given in zoccali et al .
( 2008 ) . to briefly summarize
, fibers were placed on k giants approximately 12 magnitudes brighter than the bulge red clump , and the spectra were obtained in high resolution mode ( r@xmath6@xmath16/@xmath17@xmath16@xmath120,000 ) .
the original program by zoccali et al . ( 2008 ) included four fields centered at ( l , b)=(@xmath01.14,4.18 ) , ( @xmath00.21,6.02 ) , ( 0,12 ) , and ( @xmath05.25,3.02 ) .
while the ( @xmath01.14,4.18 ) and ( @xmath00.21,6.02 ) fields were observed in the hr 13 , hr 14 , and hr 15 setups ( spanning @xmath161006950 ) , the ( @xmath05.25,3.02 ) and ( 0,12 ) fields were observed in the hr 11 , hr 13 , and hr 15 setups ( 55905835 ; 61006400 ; 66006950 ) . since the hr 11 setup is the only one containing measurable copper lines , we have only analyzed giraffe spectra from the ( @xmath05.25,3.02 ) and ( 0,12 ) fields .
we note that the ( @xmath05.25,3.02 ) field also includes the bulge globular cluster ngc 6553 .
figure [ f1 ] shows a 2mass ( skrutskie et al .
2006 ) color magnitude diagram of our final target selection from the archival data .
the raw data set obtained from the eso archive included observations of 205 rgb stars in the ( @xmath05.25,3.02 ) field and 109 rgb stars in the ( 0,12 ) field . however , we only analyzed spectra for which the co added signal to noise ( s / n ) ratio exceeded @xmath170 .
we also discarded spectra that exhibited strong tio absorption bands , for which a standard " equivalent width ( ew ) analysis would be inappropriate .
the final sample utilized here includes 75/205 stars ( 37@xmath15 ) in the ( @xmath05.25,3.02 ) field and 81/109 stars ( 74@xmath15 ) in the ( 0,12 ) field . in figure [ f1 ]
we also identify stars that are likely members of ngc 6553 ( see section 3.5 ) . in particular , note the broad dispersion in the color magnitude diagram of cluster members , as well as with stars within 5@xmath18 of the cluster center .
this highlights the combined effects of differential reddening and population mixing along the line of sight toward the ( @xmath05.25,3.02 ) field .
the star names and coordinates from the raw image headers and zoccali et al .
( 2008 ) , as well as available 2mass photometry and star identifiers , are provided in table 1 .
the raw science and calibration data were downloaded and re reduced using the giraffe base
line data reduction software ( girbldrs ) .
in particular , the pipeline software was used to carry out bias subtraction and overscan trimming , dark correction , fiber identification , flat fielding , wavelength calibration , scattered light correction , and spectrum extraction .
sky subtraction was carried out using the iraf _ skysub _ routine .
individual exposures were continuum normalized using a low order polynomial via the iraf _ continuum _ routine , and the telluric band in the hr 13 spectra was removed using the iraf task _ telluric _ and a set of flames templates obtained during a different observing program with the same spectrograph setup . the individual spectra were shifted to a common velocity scale ( i.e. , the heliocentric velocity was removed ) and co added using iraf s _ scombine _ task .
the four primary model atmosphere input parameters of effective temperature ( t@xmath19 ) , surface gravity ( log(g ) ) , metallicity ( [ fe / h ] ) , and microturbulence ( vt ) were determined via spectroscopic analyses . for stars in the ( 0,12 ) field
, we used the model parameters given in zoccali et al .
( 2008 ) as a starting point before converging to a solution .
however , the adopted model atmosphere parameters for stars in the ( @xmath05.25,3.02 ) field are not provided in zoccali et al .
( 2008 ) nor gonzalez et al .
therefore , we adopted the generic values t@xmath19=4500 k , log(g)[email protected] cgs , [ fe / h]=0.20 dex , and vt=1.5 km s@xmath20 before converging to a solution .
the final parameters given in table 1 were derived by enforcing fe i excitation equilibrium for t@xmath19 , ionization equilibrium between fe i / ii / log(g ) combination for other stars of comparable t@xmath19 and [ fe / h ] . ] for log(g ) , and removing trends in fe i abundance versus line strength for vt .
the final models were interpolated within the available grid of aodfnew ( @xmath3enhanced ) and odfnew ( scaled solar ) atlas9 model atmospheres ( castelli & kurucz 2004 ) .
stars with [ @xmath3/fe]@[email protected] were measured using the @xmath3enhanced models , and we used the scaled
solar models for stars with [ @xmath3/fe]@[email protected] .
however , the issue of an @xmath3enhanced versus scaled solar model should not introduce an error in the abundance ratios that exceeds the @xmath10.050.10 dex level ( e.g. , fulbright et al .
2006 ; alves brito et al . 2010
; johnson et al .
2013a ) .
figure [ f2 ] shows our spectroscopically determined temperature and surface gravity values in comparison with the the spectroscopic t@xmath19 and photometric log(g ) values given in zoccali et al .
as is evident in figure [ f2 ] , the spectroscopic determination of both parameters leads to a more extended distribution of surface gravities .
this is likely due to the unavoidable problem that one must assume a distance ( and mass ) when deriving a photometric surface gravity .
however , this is only a major issue when determining abundances of elements from transitions that are strongly sensitive to log(g ) .
the model atmosphere parameters determined here are well bounded by and follow the expected trends of the 10 gyr isochrones with [ fe / h]=1.5 ( @xmath3enhanced ) and [ fe / h][email protected] ( [ @xmath3/fe]=0 ) shown in figure [ f2 ] .
we do note that 25/156 ( 16@xmath15 ) stars in our sample converged to a solution in which log(g)@xmath433.5 .
the derived higher gravity values suggest some of these stars may be foreground lower rgb and subgiants rather than more evolved bulge rgb stars . a better measurement of surface gravity , either from the addition of more than 23 fe ii lines or the inclusion of more sensitive atmospheric pressure indicators , would better constrain the true nature of these stars .
we do not find any strong systematic differences in the derived [ x / fe ] ratios between stars of low " and high " gravity ratios may be an exception as the high gravity stars tend to have lower [ na / fe ] , on average .
] , but it is unclear if the similar abundances should have any bearing when interpreting bulge versus thin / thick disk composition differences ( see section 4 )
. however , the high gravity stars are also relatively metal
rich @xmath21[fe / h]@[email protected] ( @xmath23=0.25 ) , located preferentially on the blue half of the color
magnitude diagrams , and have a relatively small velocity dispersion ( @xmath23=55 km s@xmath20 for log(g)@xmath43 ) . these data provide additional circumstantial evidence that the high gravity stars may be foreground , though possibly inner disk , contaminators ( see also zoccali et al . 2008 ; their section 7 ) . in figure [ f3 ]
we compare the derived model atmosphere parameters between this study , zoccali et al .
( 2008 ) , and gonzalez et al .
we find good agreement in the derived t@xmath19 values with an average difference of only 2 k ( @xmath23=98 k ) .
the dispersion of @xmath1100 k is reasonable given the different line lists and model atmospheres ( but similar technique of excitation equilibrium ) used . as mentioned previously
, there is some discrepancy in log(g ) , especially for the highest gravity stars , between the present work and zoccali et al .
for stars with log(g)@[email protected] , the average difference in log(g ) is 0.01 dex ( @xmath23=0.29 dex ) , but for stars with log(g)@[email protected] the magnitude of the average gravity difference is 0.64 dex ( @xmath23=0.37 dex ) . comparing the microturbulence values , which may be particularly sensitive to line choice and can vary as a function of gravity , we find an average difference of 0.18 km s@xmath20 ( @xmath23=0.27 km s@xmath20 ) .
when comparing derived [ fe / h ] values , we find good agreement for [ fe / h]@[email protected] with an average difference of 0.03 dex ( @xmath23=0.13 dex ) .
however , as is evident in figure [ f3 ] our derived [ fe / h ] values are systematically higher on average by 0.18 dex ( @xmath23=0.13 dex ) , for stars with [ fe / h]@[email protected] .
the source of this discrepancy may be related to the large 1@xmath23 [ fe / h ] uncertainties given in zoccali et al .
( 2008 ) for stars with [ fe / h]@xmath20 .
this is illustrated in figure [ f4 ] where we plot the 1@xmath23 [ fe / h ] uncertainties between our work and zoccali et al .
( 2008 ) as a function of [ fe / h ] .
ideally one expects to have measurement errors that are not correlated with metallicity , as is the case here . for the zoccali et al .
( 2008 ) subsample in common with the present analysis , the line to line dispersions are comparable only for stars with [ fe / h]@[email protected] .
the abundances of fe i , fe ii , si i , ca i , cr i , and ni i were determined by measuring ews via an interactive , semi automatic code developed for this project .
the measurement process followed the standard " procedure of fitting single or multiple gaussian profiles to the spectra for isolated and weakly blended lines , respectively .
however , the measurement time frame was significantly reduced by implementing a simple machine learning algorithm that kept track of user input on a per line basis to make an educated first guess for subsequent measurements in other stars of : the number of profiles to fit , profile fitting edges , and the central wavelength , width , and central depth of all associated nearby features . while all ew measurements were manually inspected , as was mentioned in section 2 we selected stars from the archival data based primarily on s / n considerations in an effort to reduce measurement uncertainties .
sample spectra for stars of similar temperature but different metallicity are shown in figure [ f5 ] to illustrate typical data quality in the three spectrograph setups .
the line lists for this project were created by visually examining the high s / n spectra of cool metal poor and metal rich giants in the sample , finding all isolated and/or weakly blended features for elements of interest , and merging the two line list sets .
this was done to ensure that a roughly equivalent number of lines could be used in metal rich and metal poor stars , and the manual inspection of each fit enabled us to discard prohibitively strong and weak lines . on average ,
the fe i , fe ii , si i , ca i , cr i , and ni i abundances were based on the measurement of 70 , 2 , 8 , 6 , 6 , and 16 lines , respectively .
the log(gf ) values were set via an inverse abundance analysis relative to arcturus .
we adopted the arcturus model atmosphere parameters from fulbright et al .
similarly , for fe , si , and ca we adopted the arcturus abundances from fulbright et al .
( 2006 ) , and for cr and ni we adopted the arcturus abundances from ramrez & allende prieto ( 2011 ) . the final line list , including the adopted arcturus and derived solar abundances ( based on measurements of the hinkle et al . 2000 arcturus and solar atlases ) ,
are provided in table 2 .
the derived solar abundances for fe , si , ca , and cr agree within @xmath10.05 dex of the values given in asplund et al .
( 2009 ) .
the final abundances of fe i , fe ii , si i , ca i , cr i , and ni i , determined using the _ abfind _ driver of the lte line analysis code moog ( sneden 1973 ; 2010 version ) , are given in table 3 .
note also that the [ fe / h ] values given in table 1 are the average of the [ fe i / h ] and [ fe ii / h ] abundances given in table 3 .
however , the average difference in the sense [ fe i / h][fe ii / h ] is @xmath00.00 dex with a small dispersion ( @xmath23=0.02 dex ) . for the element abundances derived from transitions involving a small number of lines that are affected by significant blends from prevalent spectral features , such as molecules and ca i autoionization , and/or broadened due to isotopes and/or hyperfine structure , we used spectrum synthesis rather than ew analyses . for the present work this list includes [ o i ] , na i , mg i
, al i , co i , and cu i. the abundances were determined using the parallelized version of the _ synth _ driver for moog ( johnson et al .
2012 ) . for o ,
na , mg , and al , we adopted as a reference point the arcturus abundances given in fulbright et al .
however , as described below the reference arcturus abundances for co and cu are based on measurements using the kurucz ( 1994 ) and cunha et al .
( 2002 ) hyperfine structure line lists .
the specific reasons for using synthesis are slightly different for each element given above .
the 6300.30 [ o i ] line is blended with both a sc ii feature at 6300.69 and a ni
i feature at 6300.33 . additionally , for most stars in this sample the oxygen abundance is sensitive to the molecular equilibrium calculations set by the carbon and nitrogen abundances as well . using the cn line list from the kurucz ( 1994 ) database , we iteratively solved for the o and c@xmath0n abundances in each star . for sodium , the 6154.23 na
i line is relatively clean , but the 6160.75 na i line is partially blended with two relatively strong ca i lines .
the three mg i lines at 6319 are strongly affected by a broad ca i autoionization feature , which we set by fitting the slope of the pseudo continuum from @xmath163166318 .
the 6696.02 and 6698.67 al i lines are both affected by cn , particularly in cooler and more metal rich stars . therefore , as with [ o i ] we simultaneously fit the al i doublet and nearby cn features .
the odd z isotope @xmath24co constitutes almost 100@xmath15 of the cobalt abundance .
while the 5647.23 and 6117.00 co i lines are relatively weak ( ew @xmath1350 m ) , we included the hyperfine structure components from the kurucz ( 1994 ) line list in our syntheses . for copper , which is dominated by the two odd z isotopes @xmath25cu and @xmath26cu
, we assumed a solar system mixture of 69.17@xmath15 and 30.83@xmath15 , respectively .
we adopted the hyperfine line list of cunha et al .
( 2002 ) and derive a similar solar abundance of log @xmath11(cu)[email protected] but a slightly lower arcturus abundance than mcwilliam et al .
although the 5782.11 cu i line is strong in most of our stars , the hyperfine broadening helps desaturate the line profile to some extent .
in figure [ f6 ] we show sample syntheses of the o , mg , and cu features for a typical metal rich spectrum .
we note that the 5782 cu i line is also sometimes affected by a nearby diffuse interstellar band ( dib ) .
the width and depth of the dib feature was found to be highly variable .
the level of contamination depends on the relative velocity between the interstellar cloud and the individual star and also the reddening value .
therefore , stars in the ( @xmath05.25,3.02 ) field , which have an average e(b v)=0.7 , were more strongly affected than those in the ( 0,12 ) field , which have an average e(b v)=0.2 ( zoccali et al .
most of the stars listed in table 3 that do not have a [ cu / fe ] abundance listed were omitted because of obvious contamination by the dib feature .
radial velocities were measured using the xcsao code ( kurtz & mink 1998 ) for each individual exposure of every star and in all three filters . for reference templates we generated synthetic spectra ranging in temperature from 4250 to 5000 k ( 250 k steps ) , log(g ) from @xmath00.5 to @xmath03.5 cgs ( 0.5 dex steps ) , [ fe / h ] from 1.5 to @xmath00.5 dex ( 0.5 dex steps ) , and vt from 1 to 2 km s@xmath20 ( 0.25 km s@xmath20 steps ) .
radial velocities were determined relative to the nearest template .
we found the average agreement between exposures to be 0.15 km s@xmath20 ( @xmath23=0.13 km s@xmath20 ) .
the heliocentric corrections were taken from the headers of the pipeline reduced files , and the heliocentric radial velocities ( rv@xmath27 ) listed in table 1 represent the average value of all exposures and filters for each star . the kinematic properties of the bulge have been extensively discussed in dedicated survey papers ( e.g. , rich et al .
2007a ; howard et al .
2009 ; rangwala et al .
2009a ; babusiaux et al .
2010 ; kunder et al . 2012 ; ness et al . 2013b ; nidever et al .
2012 ; babusiaux et al .
2014 ; zoccali et al . 2014 ) .
therefore , here we seek only to place our results in context with those surveys .
figure [ f7 ] shows velocity histograms for both fields , the velocity distribution as a function of [ fe / h ] , and the velocity dispersion as a function of [ fe / h ] . while a detailed comparison between our measured velocities and those in babusiaux et al .
( 2010 ) is not possible because their individual velocities were not published , for both fields we can compare our average results with those given in figure 13 of zoccali et al .
( 2008 ) and table 3 of babusiaux et al .
for the ( @xmath05.25,3.02 ) field , ignoring ngc 6553 stars , we find average velocity and dispersion values of @xmath04.55 km s@xmath20 and 95.51 km s@xmath20 , respectively .
this compares well with the zoccali et al .
( 2008 ) average velocity of @xmath011 km s@xmath20 and velocity dispersion of 107 km s@xmath20 .
similarly , in the ( 0,12 ) field we measured an average heliocentric radial velocity of 8.61 km s@xmath20 ( @xmath23=85.56 km s@xmath20 ) compared to the babusiaux et al .
( 2010 ) values of 14 km s@xmath20 ( @xmath23=80 km s@xmath20 ) . additionally , as can be seen in figure [ f8 ] our galactocentric radial velocity ( v@xmath28 ) distributions are similar to those of nearby fields from the brava , gibs , and apogee surveys . with the exception of the stars obviously related to ngc 6553
, we find in agreement with previous bulge studies that , at least away from the galactic plane , the velocity distributions are normal with no evidence for significant cold populations ( but see also rangwala et al .
this contrasts with nidever et al .
( 2012 ) and babusiaux et al .
( 2014 ) , which find kinematically cold populations with v@xmath28@xmath1@xmath0200 km s@xmath20 . however , their fields are significantly closer to the galactic plane than those analyzed here .
we do note however that these high velocity populations are also not found in the brava , argos , nor gibs analyses , nor is there yet a satisfactory theoretical explanation for their origin ( e.g. , li et al .
2014 ) . for the ( 0,12 ) field
we observe the same trend of a decrease in velocity dispersion with increasing [ fe / h ] found by babusiaux et al .
however , while babusiaux et al .
( 2010 ) find an increase in velocity dispersion with increasing [ fe / h ] in the ( @xmath01.1,4 ) field of baade s window , our off axis but similar galactic latitude field at ( @xmath05.25,3.02 ) still exhibits a trend of decreasing velocity dispersion with increasing [ fe / h ] .
this further contrasts with recent fields observed close to the plane by babusiaux et al .
( 2014 ; see their figure 16 ) that also show a possible increase in velocity dispersion with increasing [ fe / h ] .
our result is more similar to studies of outer bulge fields that find a consistent decrease in velocity dispersion with increasing [ fe / h ] ( e.g. , johnson et al .
2011 ; uttenthaler et al .
2012 ; johnson et al . 2013 ; ness et al .
there is weak evidence in figure [ f7 ] that the trend in velocity dispersion and [ fe / h ] may be more shallow for the ( @xmath05.25,3.02 ) field compared to the ( 0,12 ) field .
note that the inclusion ( or not ) of ngc 6553 stars significantly affects the velocity dispersion of the [ fe / h ] bin in which the cluster resides .
members of the globular cluster ngc 6553 in the ( @xmath05.25,3.02 ) field are best identified in the velocity metallicity diagram in figure [ f7 ] .
the likely members ( 12 stars total ) are clustered near [ fe / h]@xmath290.10 and rv@xmath27@xmath290 km s@xmath20 .
literature measurements of the cluster s average [ fe / h ] value vary considerably , with estimates that include : 0.55 ( barbuy et al .
1999 ) , 0.16 ( cohen et al .
1999 ) , 0.7 ( coelho et al . 2001 ) , 0.3 ( origlia et al .
2002 ) , 0.2 ( melndez et al .
2003 ) , and 0.2 ( alves brito et al .
however , we find in agreement with the most recent estimates that @xmath21[fe / h]@xmath22=0.11 ( @xmath23=0.07 ) . while the cluster is slightly iron
deficient relative to the sun , the moderate enhancements of the cluster s [ @xmath3/fe ] ratio ( see section 4.1 ) gives it an overall metallicity that is roughly solar .
ngc 6553 is therefore one of the most metal rich globular clusters in the galaxy .
we find similar agreement with literature values for the cluster s radial velocity , with @xmath21rv@xmath27@xmath22=2.03 km s@xmath20 ( @xmath23=4.85 km s@xmath20 ) .
this is compared with recent values of : 1 km s@xmath20 ( coelho et al .
2001 ) , @xmath01.6 km s@xmath20 ( melndez et al .
2003 ) , and 1.86 km s@xmath20 ( alves brito et al .
finally , we note that the stars identified in table 1 as possible cluster members have an average , projected radial distance from the cluster center of about 6@xmath18 ( @xmath23=5@xmath18 ) . we have adopted a more lenient radial distance discriminator than the 2@xmath18 limit used by zoccali et al .
( 2008 ) and gonzalez et al .
( 2011 ) , and instead rely more on the [ fe / h ] and velocity measurements to identify possible cluster members . as noted previously , zoccali et al .
( 2008 ) and gonzalez et al .
( 2011 ) presented [ fe / h ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] abundances based on the same giraffe data utilized here .
therefore , in figures [ f9][f10 ] we compare our results with theirs for stars and elements in common . while a quantitative comparison of the individual [ fe / h ] values is given in section 3.1 ( see also figure [ f3 ] ) , in figure [ f9 ] we compare the general shapes and bulk properties of the metallicity distribution functions . for the ( @xmath05.25,3.02 )
field the average and median [ fe / h ] ratios are similar , but the distribution from the present work is somewhat broader and extends to higher [ fe / h ] . in contrast , there are no significant differences in the [ fe / h ] distribution functions between the present work and the same stars from zoccali et al .
( 2008 ) , for the ( 0,12 ) field .
we also reconfirm one of the primary conclusions of zoccali et al .
( 2008 ) that interior bulge fields have a higher average metallicity than outer bulge fields .
finally , we note that the distribution functions shown in figure [ f9 ] do not provide strong evidence supporting the existence of multiple , discreet populations , as has been suggested in some studies ( bensby et al .
2011 ; hill et al .
2011 ; bensby et al . 2013 ; ness et al .
however , the number of stars per field presented here is @xmath14100 . in figure [ f10 ]
we compare our derived [ mg / fe ] , [ si / fe ] , and [ ca / fe ] ratios to those given in gonzalez et al .
the average differences between the present work and that of gonzalez et al .
( 2011 ) are @xmath17[mg / fe][email protected] ( @xmath23=0.14 ) , @xmath17[si / fe][email protected] ( @xmath23=0.13 ) , and @xmath17[ca / fe][email protected] ( @xmath23=0.14 ) . the relatively consistent star to star scatter of @xmath10.14 dex is a reasonable estimate of the attainable precision between the two studies , which derive @xmath3element abundances from different techniques ( synthesis in gonzalez et al .
2011 and ew measurements here ) .
we note that the @xmath3elements oxygen ( measured here ) and titanium ( measured in gonzalez et al .
2011 ) were not both measured in each study .
feature because of concerns regarding measurement accuracy .
we chose not to include ti abundances because of discrepant nucleosynthesis predictions for this element in comparison to observations ( e.g. , see kobayashi et al .
2011 ; their figure 14 ) . ]
we investigated the sensitivity of derived abundances for each element in every star by taking the abundances given in table 3 , determining theoretical ews using the line list in table 2 , and then varying the model atmosphere parameters t@xmath19 , log(g ) , [ fe / h ] , and vt individually while holding the other parameters fixed .
we selected parameter changes of 100 k in t@xmath19 , 0.30 dex in log(g ) , 0.15 dex in [ m / h ] , and 0.30 km s@xmath20 in vt , which are reasonable when comparing our derived parameters with those of the independent analysis by zoccali et al .
( 2008 ; see also section 3.1 ) .
the total uncertainty for each element ratio in each star resulting from this exercise is provided in table 4 . in general , most elements are not affected by changes in t@xmath19 of 100 k at more than the 0.1 dex level .
however , the two species presented here that reside in their dominant ionization states ( [ o i ] and fe ii ) are strongly affected by changes in surface gravity . for a change in log(g ) of 0.3 dex , the log @xmath11(o ) and log @xmath11(fe ii ) abundances can change by @xmath10.10.3 dex , but these effects are mitigated when the [ o i / h ] abundance is normalized with [ fe ii / h ] .
these two species are also more strongly affected by changes in the model metallicity , and the larger [ fe ii / h ] measurement and sensitivity uncertainties are a contributing factor to the increased dispersion in the [ o / fe ] ratios compared to other @xmath3elements ( e.g. , [ mg / fe ] ) . as expected , microturbulence sensitivity is correlated with a star s overall metallicity ( i.e. , line strength ) . among the transitions under consideration here , in metal rich stars those of na , ca , and cu typically have the strongest lines and are thus more strongly affected by the microturbulence uncertainty . in table 5
we also provide the 1@xmath23 line to line dispersion values for all species measured here .
these values should be mostly representative of the combined measurement error that includes effects such as : continuum placement , line deblending , synthesis fits via visual inspection , log(gf ) uncertainties , and model atmosphere deficiencies .
typical line to line dispersion values are @xmath10.08 dex .
the measurement error of cu may be underestimated because of the line s large ew , non negligible blending ( see figure [ f6 ] ) , and possible contamination with a nearby dibs feature .
a more reasonable measurement uncertainty for cu is , in most cases , @xmath10.150.20 dex .
the @xmath3elements have been the primary focus of detailed composition work in the galactic bulge . to first order
there is agreement among the various studies that : ( 1 ) the [ @xmath3/fe ] ratios are enhanced by @[email protected] dex at [ fe / h]@xmath130.3 , ( 2 ) for stars with [ fe / h]@[email protected] there is a mostly monotonic decline in [ @xmath3/fe ] with increasing [ fe / h ] , ( 3 ) the bulge and thick disk may share similar chemistry over a wide range in metallicity , and ( 4 ) there are no significant variations in the [ @xmath3/fe ] trends between different bulge sight lines ( mcwilliam & rich 1994 ; cunha & smith 2006 ; zoccali et al .
2006 ; fulbright et al . 2007 ; lecureur et al . 2007 ; melndez et al . 2008 ; alves brito et al . 2010 ; bensby et al . 2010a ; ryde et al . 2010 ; bensby et al . 2011 ; gonzalez et al . 2011 ; hill et al . 2011 ; johnson et al . 2011 ; uttenthaler et al .
2012 ; bensby et al . 2013 ; garca prez et al .
2013 ; johnson et al . 2013 ; ness et al . 2013a ; jnsson et al . 2014 ) .
additionally , there is evidence that the [ o / mg ] ratio declines with increasing metallicity ( fulbright et al . 2007 ; lecureur et al .
2007 ; mcwilliam et al . 2008 ; alves brito et al . 2010 ) .
the new data presented here , and summarized in figure [ f11 ] , reinforce many observations from the previous studies mentioned above .
in particular , we find that for [ fe / h]@xmath140.3 all of the [ @xmath3/fe ] ratios are enhanced and exhibit minimal star to star scatter with @xmath21[o / fe]@[email protected] ( @xmath23=0.10 ) , @xmath21[mg / fe]@[email protected] ( @xmath23=0.08 ) , @xmath21[si / fe]@[email protected] ( @xmath23=0.07 ) , and @xmath21[ca / fe]@[email protected] ( @xmath23=0.09 ) . for stars with [ fe / h]@xmath40.3 , we find that the [ @xmath3/fe ] ratios decrease with increasing [ fe / h ]
. however , figure [ f11 ] illustrates the disparate trends for individual elements and highlights the information loss that can occur when averaging abundance ratios for multiple @xmath3elements .
the [ o / fe ] ratios are higher by @xmath10.2 dex in metal poor stars than those of other @xmath3elements , but this trend reverses for stars with [ fe / h]@xmath20 where [ o / fe ] is , on average , lower by @xmath10.2 dex . while both o and mg are significant products of hydrostatic burning in massive stars ( e.g. , woosley & weaver 1995 ) , the [ mg / fe ] trend exhibits a more shallow decline with increasing [ fe / h ] than [ o / fe ] .
this is most clearly seen in figure [ f12 ] , which shows a sharply declining [ o / mg ] ratio at [ fe / h]@xmath20.1 .
although massive star production of si and ca involves both hydrostatic and explosive burning ( e.g. , woosley & weaver 1995 ) , the [ si / fe ] and [ ca / fe ] trends are more similar to [ mg / fe ] than [ o / fe ] . given the disparate trend of [ o / fe ] compared to other @xmath3elements , and the low production of most @xmath3elements relative to fe in type ia supernovae ( sne ; e.g. , nomoto et al .
1997 ) , we conclude in agreement with past work ( e.g. , mcwilliam et al . 2008 ) that the strong decline in [ o / fe ] at [ fe / h]@xmath20.3 is likely a result of metallicity dependent yields in massive stars yield ( e.g. , kobayashi et al . 2006 ; their figure 9 ) . producing models with a declining [ o / mg ]
yield likely requires the inclusion of additional physics , such as mass loss from stellar winds or a process to cause a change in the [ c / o ] ratio . ] .
the influence of metallicity dependent yields on the bulge composition profile , especially from wolf rayet stars , is further supported by fluorine measurements ( cunha et al .
2008 ; jnsson et al . 2014 ) , but it remains to be seen if this scenario can be reconciled with the observed carbon and nitrogen trends ( ryde et al .
2010 ; but see also cescutti et al .
2009 ) .
when comparing the individual [ @xmath3/fe ] trends between the two fields analyzed here , figure [ f11 ] shows no significant variations .
similarly , in figure [ f13 ] we combine our two fields and compare with literature giant and dwarf [ @xmath3/fe ] data .
a comparison between the present work and literature giant trends , which span a variety of bulge sight lines , leads us to find in agreement with johnson et al .
( 2011 ; 2013 ) and gonzalez et al . ( 2011 ) that no significant field to field [ @xmath3/fe ] variations exist over a broad region of the bulge .
the microlensed dwarf data exhibit the same qualitative and quantitative distributions for [ o / fe ] and [ mg / fe ] , at least for [ fe / h]@xmath130 , as the giant data , but there may be small systematic offsets with [ si / fe ] and [ ca / fe ] .
in particular , the dwarf abundances are @xmath10.1 dex lower for a given [ fe / h ] , when considering [ fe / h]@xmath130 . at super solar [ fe / h ] values , the dwarf and giant data are in excellent agreement for [ o / fe ] , but the leveling off or slight increase in [ mg / fe ] , [ si / fe ] , and [ ca / fe ] seems to be unique to the dwarf measurements . unfortunately , the source of this discrepancy is not clear and may be related to analysis differences between dwarfs and giants . in figure [ f14 ]
we compare the [ o / fe ] , [ mg / fe ] , [ si / fe ] , and [ ca / fe ] abundances between the bulge , thick disk , and thin disk . for stars with [ fe / h]@xmath130.5 , the bulge and thick disk stars exhibit similar abundance patterns for all four element ratios .
however , we note that on average the bulge stars have [ o / fe ] and [ mg / fe ] ratios that are slightly enhanced by @xmath10.03 dex and [ si / fe ] and [ ca / fe ] ratios that are enhanced by @xmath10.05 dex compared to similar metallicity thick disk stars . in contrast , the most metal poor thin disk stars exhibit significantly lower [ x / fe ] ratios for all of the @xmath3elements measured here .
the bulge and thin disk stars with [ fe / h]@xmath20 are not strikingly different , but the star to star scatter , especially for [ o / fe ] , is significantly larger for the bulge giants . for the intermediate range of [ fe / h]@xmath10.5 to 0 , the bulge stars still exhibit significantly larger [ @xmath3/fe ] ratios than the thin disk and may remain enhanced to a higher [ fe / h ] value than the thick disk . the chemical similarities between especially the metal poor bulge and thick disk found here have been documented in previous work ( melndez et al . 2008 ; alves brito et al .
2010 ; bensby et al . 2010a ; ryde et al . 2010
; bensby et al . 2011 ; gonzalez et al .
2011 ; hill et al .
2011 ; johnson et al .
2011 , 2013 ) . the apparent homogeneity between the most metal poor bulge and thick disk stars lends credibility to the idea that the galactic bulge formed _ in situ _ with the disk
. however , there is not universal agreement in the literature that the metal poor bulge and disk trends are identical . in particular , earlier work by zoccali et al .
( 2006 ) , fulbright et al .
( 2007 ) , and lecureur et al .
( 2007 ) found that the bulge stars exhibited both larger [ @xmath3/fe ] ratios and remained enhanced to higher [ fe / h ] values than the local thick disk , which implies a more rapid formation timescale for the bulge .
in contrast , purely differential analyses between thick disk and bulge giants ( melndez et al . 2008 ; alves brito et al . 2010 ; gonzalez et al . 2011 ) find nearly identical [ @xmath3/fe ] versus [ fe / h ] trends , at least for [ fe / h]@xmath130.3 .
however , bensby et al . (
2013 ) noted in a similarly differential comparison of local thick disk dwarfs and bulge microlensed dwarfs that the inflection point at which [ @xmath3/fe ] declines may be 0.10.2 dex higher ( [ fe / h]@xmath290.3 to 0.2 ) in the bulge .
while we compare bulge giants and thick disk dwarfs in figure [ f14 ] , our results are in agreement with bensby et al .
in particular , we find that the [ mg / fe ] , [ si / fe ] , and perhaps [ o / fe ] ratios remain enhanced to a higher [ fe / h ] value than those of the local thick disk / fe ] ratios between bulge giants here and thick disk giants from alves
brito et al .
( 2010 ) , we reach a similar conclusion . both data sets exhibit similar abundance trends for [ o / fe ] and [ mg / fe ] , but
[ si / fe ] and [ ca / fe ] remain enhanced at higher [ fe / h ] in the bulge giants . ] .
finally , we note that combining the present data set with those available in the literature ( e.g. , see figure [ f13 ] ) totals @xmath110@xmath31 bulge stars that have had [ @xmath3/fe ] measurements made from high resolution spectroscopy . despite the large sample size , there is a paucity of stars with [ @xmath3/fe ] ratios that deviate significantly from the bulk trend . in agreement with work suggesting
the galactic bulge did not form predominantly from a build up of merger events ( e.g. , shen et al .
2010 ) , we can effectively rule out significant contributions from the infall of objects with chemistry similar to those of many present day dwarf galaxies ( i.e. , low [ @xmath3/fe ] ; e.g. , see venn et al .
2004 and references therein ) . additionally , as can be seen in figure [ f11 ]
( see also gonzalez et al .
2011 ) the [ x / fe ] abundance ratios of individual @xmath3elements for ngc 6553 stars are nearly identical to those of bulge field stars with similar [ fe / h ] .
specifically , the average [ x / fe ] values for ngc 6553 are : @xmath21[o / fe]@[email protected] ( one star ) , @xmath21[mg / fe]@[email protected] ( @xmath23=0.08 ) , @xmath21[si / fe]@[email protected] ( @xmath23=0.10 ) , and @xmath21[ca / fe]@[email protected] ( @xmath23=0.12 ) , which compare well with the average abundances for nearby bulge field stars in the range [ fe / h]=0.20 to @xmath00.00 : @xmath21[o / fe]@[email protected] ( @xmath23=0.29 ) , @xmath21[mg / fe]@[email protected] ( @xmath23=0.09 ) , @xmath21[si / fe]@[email protected] ( @xmath23=0.08 ) , and @xmath21[ca / fe]@[email protected] ( @xmath23=0.13 ) .
these values are in good agreement with past work that finds the cluster to be moderately @xmath3enhanced ( barbuy et al .
1999 ; cohen et al .
1999 ; coelho et al . 2001
; origlia et al . 2002 ; melndez et al . 2003 ; alves brito et al .
the similar [ @xmath3/fe ] abundances between the cluster and field stars suggests that ngc 6553 likely formed _ in situ _ with the bulge field population and is not a captured cluster . in a similar fashion to the @xmath3elements , the light , odd z elements na and al provide clues of the processes that dominated the chemical enrichment of a stellar population .
furthermore , these elements are useful for chemical tagging " analyses , and both the [ na / fe ] and [ al / fe ] ratios can vary significantly between stellar populations that have otherwise identical [ @xmath3/fe ] and [ fe / h ] values . the large ( @xmath20.5 dex ) star to star [ na / fe ] and [ al / fe ] abundance variations present in metal poor globular cluster but not halo / disk stars of the same metallicity are perhaps the most well known example of this phenomenon ( e.g. , see reviews by gratton et al . 2004 ; 2012 and references therein ) .
while the production of na and al is dominated by hydrostatic helium , carbon , and neon burning in massive stars , the final yields are expected to grow significantly with increasing progenitor mass and metallicity ( e.g. , woosley & weaver 1995 ; kobayashi et al . 2006 ; 2011 ) .
intermediate mass ( @xmath148 m@xmath5 ) asymptotic giant branch ( agb ) stars and the hydrogen rich envelopes of massive stars can also produce significant amounts of na and al via the nena and mgal proton capture cycles ( e.g. , decressin et al . 2007 ; de mink et al .
2009 ; ventura & dantona 2009 ; karakas 2010 ) .
since na and al are thought to result from similar production mechanisms , we expect their abundance patterns to reflect a comparable morphology . while the bulge abundance patterns of [ na / fe ] and [ al / fe ] have not been investigated to the extent of the @xmath3elements , the combined literature sample now totals of order a few hundred stars .
interestingly , the agreement between studies regarding the [ na / fe ] and [ al / fe ] trends is worse than for the @xmath3elements . while all high resolution analyses ( mcwilliam & rich 1994 ; fulbright et al .
2007 ; lecureur et al . 2007 ; alves brito et al . 2010 ; bensby et al . 2010a , 2011 ; johnson et al .
2012 ; bensby et al .
2013 ) tend to agree that the average [ na / fe ] ratio rises with increasing metallicity , significant scatter is present at [ fe / h]@xmath131 and [ fe / h]@xmath20 .
similarly , there is general agreement that [ al / fe ] is enhanced in bulge stars at [ fe / h]@xmath130.3 .
however , some studies find that [ al / fe ] remains enhanced at super solar metallicities ( mcwilliam & rich 1994 ; fulbright et al . 2007 ; lecureur et al .
2007 ; alves brito et al .
2010 ) while others find a decline in [ al / fe ] , similar to [ @xmath3/fe ] ( bensby et al . 2011 , 2011 ; johnson et al . 2012 ; bensby et al . 2013 ) .
additionally , there is general agreement that the [ na / fe ] and [ al / fe ] trends as a function of [ fe / h ] are similar between the bulge and disk over a broad metallicity range , but differences could be present at the metal poor and metal rich ends of the bulge distribution .
it is also not yet clear if any significant [ na / fe ] and [ al / fe ] abundance differences exist between different bulge sight lines .
figure [ f11 ] shows our derived [ na / fe ] and
[ al / fe ] abundances as a function of [ fe / h ] for both fields and the possible ngc 6553 stars , and in figure [ f15 ] we compare our results with those from previous work . for na
we find general agreement with literature values such that the average [ na / fe ] ratio rises with increasing [ fe / h ] .
however , we find only a small number of metal rich stars with [ na / fe]@[email protected] and do not reproduce the very large [ na / fe ] ratios of lecureur et al .
additionally , we do not find significant evidence supporting large [ na / fe ] variations between the two bulge sight lines probed here . at [ fe / h]@xmath20.5 , the mean [ na / fe ] trend and star to
star dispersion for our measured rgb stars is in good agreement with those of the microlensed bulge dwarfs ( e.g. , bensby et al .
2013 ) . the primary discrepancy between our work and some of the literature values occurs for stars with [ fe / h]@xmath130.7 , with the present work and johnson et al .
( 2012 ) finding that the average na trend decreases from [ na / fe]@xmath10 at [ fe / h]=0.5 to [ na / fe]=0.3 at [ fe / h]=1.7 .
it is not immediately clear if the discrepancy , especially between the bulge rgb and dwarf data , is real or caused by analysis differences ( e.g. , nlte , 3d , or spherical / plane parallel effects between dwarfs and giants ) .
the inclusion of nlte corrections would minimize the differences at low metallicity between bulge rgb and dwarf stars , and also between bulge rgb and metal poor thick disk dwarfs ( see figure [ f16 ] ) , if the largely positive na corrections for rgb stars from gratton et al .
( 1999 ) were applied . however , more recent nlte calculations ( e.g. , lind et al .
2011 ) instead find that the sign of the na correction is negative for the lines and atmospheric parameters used here .
similarly , the nlte corrections for log @xmath11(fe i ) appear to be positive ( e.g. , lind et al .
2012 ; bergemann et al .
2012 ) for most stars in our sample , which would decrease the [ na / fe ] ratios .
further insights into this problem may be gained as more extensive nlte calculations and 3d model atmosphere grids and codes become available .
when comparing the [ na / fe ] and [ al / fe ] trends in figure [ f11 ] , it is immediately clear that the two elements exhibit discrepant trends .
while [ na / fe ] gradually rises with increasing [ fe / h ] , the [ al / fe ] trend is nearly indistinguishable from that of most @xmath3elements . in particular
, we find in agreement with bensby et al .
( 2010a ; 2011 ; 2013 ) and johnson et al .
( 2012 ) that [ al / fe]@[email protected] in bulge stars until [ fe / h]@xmath10.3 and then steadily declines at higher [ fe / h ] . as mentioned previously , the decline in [ al / fe ] with increasing metallicity contrasts with other literature results that find [ al / fe ] remains enhanced even at [ fe / h][email protected] ( fulbright et al . 2007 ; lecureur et al . 2007 ; alves
brito et al . 2010 ) .
the data quality among the various studies is comparable , and it is not clear why the derived [ al / fe ] trends are in disagreement at high metallicity .
we do note however that for cool , high metallicity stars the 6696 and especially 6698 al i lines , as well as the continuum placement , can be affected by cn blending .
the discrepant [ na / fe ] and [ al / fe ] trends as a function of [ fe / h ] are not limited to the bulge and may also be present in the disk , as can be seen in figure [ f16 ] . despite nucleosynthesis models predicting similar production of na and al in massive stars ( e.g. , woosley & weaver 1995 ) , figure [ f16 ]
shows that , at least in the metallicity range probed here , al is over produced relative to na in both bulge and disk stars for [ fe / h]@xmath130.3 .
the increased production of na relative to al in metal rich stars , and especially in the bulge , suggests that metallicity dependent yields from massive stars vary more strongly for na than al . contributions from intermediate mass agb stars
may also help explain the na and al trends since the agb [ na / fe ] yields tend to increase at higher [ fe / h ] while those of [ al / fe ] decline ( e.g. , ventra & dantona 2009 ) .
interestingly , we find that , unlike the case for [ na / fe ] , the [ al / fe ] ratios are nearly indistinguishable between the bulge and thick disk at [ fe / h]@xmath140 .
similarly , the [ al / fe ] ratios for bulge stars are identical to those in the thin disk at [ fe / h]@xmath40 . given the similar behavior of [ al / fe ] to many of the @xmath3elements , in figure [ f17 ] we provide a detailed comparison between [ al / fe ] , [ o / fe ] , [ mg / fe ] , [ si / fe ] , and [ ca / fe ] for the bulge stars analyzed here . while the [ o / fe ] trend is clearly different than that of [ al / fe ] , there are no similarly strong discrepancies between [ al / fe ] and the other @xmath3elements . at [ fe / h]@xmath140.8 both [ mg / fe ] and
[ ca / fe ] are @xmath10.100.15 dex enhanced compared to [ al / fe ] , but those differences disappear at higher [ fe / h ] . on the other hand ,
the [ si / fe ] and [ al / fe ] trends are essentially identical at all [ fe / h ] with an average difference of 0.01 dex ( @xmath23=0.13 dex ) . examining the ngc 6553 stars in figure [ f11 ] shows that na , and to a lesser extent al , exhibit larger star to star [ na / fe ] and [ al / fe ] variations than similar metallicity field stars
. in particular , the average na and al abundances for the cluster stars are @xmath21[na / fe]@[email protected] ( @xmath23=0.20 ) and @xmath21[al / fe]@[email protected] ( @xmath23=0.13 ) , which can be compared to the similar metallicity fields stars having @xmath21[na / fe]@[email protected] ( @xmath23=0.11 ) and @xmath21[al / fe]@[email protected] ( @xmath23=0.10 ) , respectively .
the larger [ na / fe ] abundance and dispersion values for the cluster stars suggests ngc 6553 experienced some degree of self enrichment .
however , unlike low metallicity globular clusters , ngc 6553 does not exhibit a strong na
al correlation .
this is in agreement with the observed trend that the na
al correlation is more mild and [ al / fe ] dispersions smaller in metal rich as opposed to metal poor globular clusters ( e.g. , carretta et al .
2009 ; oconnell et al . 2011 ; cordero et al . 2014 ) . unfortunately , the 6300 telluric oxygen emission feature combined with ngc 6553 s relatively low radial velocity prohibited us from obtaining an [ o / fe ] abundance for more than one star in ngc 6553 .
therefore , we can not comment further on the existence or extension of the likely o
na correlation .
finally , we note that our mean [ na / fe ] and [ al / fe ] values and abundance dispersions are in excellent agreement with those found by alves
brito et al . ( 2006 ) , but are considerably lower than the values ( based on two stars ) of barbuy et al . ( 1999 ) .
unlike the lighter elements , the abundance patterns of fe peak elements in the galactic bulge are not well explored .
the production of fe peak elements occurs through a variety of processes in the late stages of massive star evolution , the resulting core collapse sne , and also in type ia sne .
peak abundance patterns can also be useful indicators of a stellar population s imf , with odd z elements in particular providing some diagnostic power ( e.g. , nomoto et al .
some initial work on the bulge fe
peak abundance distribution was included in mcwilliam & rich ( 1994 ) , which found [ v / fe ] , [ cr / fe ] , and [ ni / fe ] ratios near solar and a possible enhancement in [ co / fe ] and [ sc / fe ] .
more recent work analyzing the fe
peak abundance trends in the bulge has come from microlensed dwarf studies ( cohen et al .
2008 ; johnson et al . 2008 ; cohen et al . 2009 ; bensby et al .
2010a ; epstein et al .
2010 ; bensby et al . 2011 ; bensby et al .
the bulge [ mn / fe ] trend in rgb stars has also been investigated recently by barbuy et al .
the results of the these analyses indicate that the bulge fe peak trends are similar to that of the local disk , except that the bulge may have different [ mn / o ] ratios than the thick disk for a given [ o / h ] value .
the general [ x / fe ] versus [ fe / h ] abundance trends derived here are shown in figure [ f11 ] . from these data
we find that : ( 1 ) cr is the element that most closely tracks fe with @xmath21[cr / fe]@xmath22=0.00 ( @xmath23=0.11 ) , ( 2 ) [ co / fe ] exhibits low level variations as a function of [ fe / h ] but is generally enhanced with @xmath21[co / fe]@[email protected] ( @xmath23=0.11 ) , ( 3 ) [ ni / fe ] shows similar variations to [ co / fe ] but at a much smaller amplitude and is slightly enhanced with @xmath21[ni / fe]@[email protected] ( @xmath23=0.06 ) , ( 4 ) the cu abundance increases monotonically from [ cu / fe]=0.84 in the most metal poor star to [ cu / fe]@[email protected] in the most metal rich stars , and ( 5 ) there are no significant fe
peak abundance variations between ngc 6553 stars and the field stars . although the exact nature of cu nucleosynthesis is complex ( e.g. , see mishenina et al .
2002 and references therein ) , the significant secondary ( i.e. , metallicity dependent ) production of cu ( and also na ) is evident in figure [ f11 ] .
additionally , figure [ f12 ] shows that despite the larger measurement errors in both o and cu abundances , the [ cu / o ] ratio is strongly correlated with [ fe / h ] .
this trend has been noted previously and is prevalent in stellar populations with different star formation histories , such as the local disk and sagittarius dwarf galaxy ( e.g. , mcwilliam et al . 2013 ) .
the [ cu / o ] trend is taken as evidence that a significant portion of cu is synthesized in massive stars , perhaps via the weak s process ( e.g. , sneden et al . 1991 ) .
however , some component of cu may also be produced by type ia sne ( matteucci et al .
1993 ) . in figure [ f18 ]
we compare our derived fe
peak abundance trends with those in the literature . for cr
there is general agreement between the bulge rgb stars analyzed here and the literature microlensed dwarf data .
however , the small number of bulge literature data points for co and cu makes a direct comparison difficult .
the [ ni / fe ] comparison also shows excellent agreement overall , but the rgb stars appear systematically enhanced by @xmath130.1 dex in the range [ fe / h]=0.3 to @xmath00.1 .
note also the similarly small star to star dispersion in especially [ ni / fe ] between the rgb and dwarf data .
a comparison between the bulge fe peak abundance trends and those of the thin / thick disk is shown in figure [ f19 ] .
interestingly , at least for [ fe / h]@xmath21.5 , the [ cr / fe ] distribution is seemingly independent of formation environment with the bulge , thick disk , and thin disk stars all having [ cr / fe]@xmath10 . for [ co / fe ] , [ ni / fe ] , and [ cu / fe ]
, there is significant overlap between the bulge and thick disk trends at [ fe / h]@xmath130.5 . at higher [ fe / h ] ,
the bulge may be enhanced in all three elements relative to both the thick and thin disks .
this is especially evident in the figure [ f19 ] panel showing [ ni / fe ] versus [ fe / h ] ; the low star to star scatter in [ ni / fe ] for all three populations highlights the possible composition difference between the local disk and bulge from [ fe / h]@xmath10.4 to @xmath00.2 . while the strong rise in [ cu / fe ] with metallicity is , as mentioned previously , a common feature in many different stellar populations , the bulge stars at [ fe / h]@xmath20.3 appear to extend to higher abundances than the local disk
. however , the increased measurement uncertainty of cu and paucity of disk [ cu / fe ] ratios at [ fe / h]@xmath40 prevents us from undertaking a more comprehensive analysis . accurately modeling the chemical enrichment history of a stellar system requires solving for a variety of free parameters that may include the imf , star formation rate , star formation efficiency , supernova / hypernova ratio20 m@xmath5 ] , inflow / outflow rate , binary fraction , stellar evolution time scales , mass loss rates , and stellar yields . while not all of the required input parameters are yet well defined based on observed data , chemical enrichment models are effective tools for examining and interpreting chemical composition data .
therefore , in figures [ f20][f21 ] we compare our derived abundance trends with those predicted by chemical enrichment models in which parameters such as the imf , binary fraction , supernova / hypernova ratio , and outflow rate are varied .
the baseline galactic bulge model shown in figures [ f20][f21 ] is from kobayashi et al .
( 2006 ; 2011 ) and is designed to reproduce the metallicity distribution in baade s window from zoccali et al .
( 2008 ) , assumes a kroupa ( 2008 ) imf , and assumes a star formation time scale of 3 gyr ( see kobayashi et al .
2011 ; their table 1 and section 2.4 for more detail regarding model input parameters ) . in general , the baseline model does a reasonable job of reproducing the observed abundance trends of all abundance ratios , except [ na / fe ] and [ al / fe ] .
all of the models shown in figures [ f20][f21 ] predict large over abundances of both [ na / fe ] and [ al / fe ] that are not observed , suggesting the massive star yields of both elements may be too high .
however , as can be seen in figure [ f20 ] , the enhanced fe production from hypernovae ( hne ) decreases the [ na / fe ] and [ al / fe ] yields and brings the baseline bulge model into better agreement with the light element data .
the addition of hne also provides better agreement between the models and observations for the fe
peak elements , with a trade off of [ @xmath3/fe ] ratios that may be slightly too low .
in contrast , figure [ f20 ] also shows that a paucity of hne generally leads to [ x / fe ] ratios that are too high .
it seems that a significant fraction of hne are required to accurately reproduce the observed bulge abundance trends .
unfortunately , the hn fraction is best constrained at [ fe / h]@xmath131 , where data are scarce . in figure [ f21 ]
we examine how changes in the imf could affect the expected abundance trends .
compared to the kroupa ( 2008 ) imf adopted in our baseline bulge model , a steep imf ( x=1.6 ) is completely ruled out by the data . additionally , adopting a kroupa ( 2008 ) imf that truncates at 40 m@xmath5 , and thus ignores contributions from the most massive stars ,
is inconsistent with the [ cu / fe ] abundances , and to a lesser extent those of [ co / fe ] . while a flatter , top
heavy imf ( x=0.3 ) alone leads to [ x / fe ] ratios that are too high for nearly every element , a reduction in the yields from outflow and/or slow star
formation combined with a high type ia sn rate , artificially enhanced with a ten times larger binary fraction , could bring such a model into agreement with the data . however , bulge formation models with slow star formation are likely unrealistic , and the observed [ co / fe ] and [ cu / fe ] data appear to rule out these models . based on the present data it does not appear that the bulge required a uniquely non standard " imf to reach its present day composition ( but see also ballero et al .
2007 , for example ) .
finally , in figure [ f20 ] we also compare the measured bulge abundance trends with our adopted baseline model and similar models representing the composition distributions of the local thick disk and thin disk . comparing the three predicted trends indicates that in the range 0.8@xmath13[fe / h]@xmath130.3 the bulge [ @xmath3/fe ] ratios should be similar or modestly enhanced and remain enhanced to higher [ fe / h ] than the thick disk .
similarly , at [ fe / h]@xmath20 the bulge and thin disk should exhibit similar , if not identical , [ @xmath3/fe ] ratios . both of these predictions match our observations ( see section 4.1 ) .
the predicted enhancements in the bulge for [ na / fe ] and [ al / fe ] compared to the local disk are not supported by observations , but this could be related to the previously mentioned possible over production issues of the adopted stellar yields . however , in addition to na and al , figure [ f20 ] shows that co and cu may also exhibit some discriminating power between the bulge and local disk populations . in particular , the data support bulge stars with [ fe / h]@xmath20.5 having [ co / fe ] and [ cu / fe ] ratios that are higher than the local disk .
therefore , the data and models presented here provide some supporting evidence that the bulge experienced a different chemical enrichment path than the thick disk .
we have measured radial velocities and chemical abundances of o , na , mg , al , si , ca , cr , fe , co , ni , and cu in a sample of 156 rgb stars located in galactic bulge fields centered near ( l , b)=(@xmath05.25,3.02 ) and ( 0,12 ) . the ( @xmath05.25,3.02 ) also includes 12 stars identified as likely members of the bulge globular cluster ngc 6553 , based on their radial velocity and [ fe / h ] values .
the results are based on high resolution archival spectra obtained with the flames giraffe instrument , and originally used to derive [ fe / h ] and [ @xmath3/fe ] abundances in zoccali et al .
( 2008 ) and gonzalez et al .
we culled the original target list and selected only those stars with co added s / n@xmath270 that also lack strong tio bands .
the abundance analysis was carried out using standard ew and spectrum synthesis techniques .
our derived heliocentric radial velocity distributions for both fields are in good agreement with past surveys ( brava , gibs , and apogee ) covering nearby fields .
we do not confirm the existence of a significant population of high velocity stars noted by nidever et al .
( 2012 ) and babusiaux et al .
. however , our targeted fields are farther away from the plane than most of those in which nidever et al .
( 2012 ) and babusiaux et al .
( 2014 ) observe the cold , high velocity stars . for both fields analyzed here
we also find that the velocity dispersion monotonically decreases with increasing [ fe / h ] .
this is not unexpected for the outer bulge field at ( 0,12 ) , but the similar trend in the ( @xmath05.25,3.02 ) field appears to contradict the findings of babusiaux et al .
( 2010 ; 2014 ) that the velocity dispersion of bulge stars with [ fe / h]@xmath20 increases at lower galactic latitude .
the reason for this discrepancy is not clear , but we note that previous analyses finding increased velocity dispersion at low galactic latitude for metal rich stars have all focused on minor axis fields . the inner bulge field included here is several degrees off axis .
the composition data reconfirm the already well documented metallicity gradient in the bulge .
similarly , we find good agreement between our derived [ mg / fe ] , [ si / fe ] , and [ ca / fe ] abundances and those of gonzalez et al . ( 2011 ) . additionally , we confirm that there are no significant field to field [ @xmath3/fe ] abundance variations among various bulge sight lines .
our new @xmath3element measurements also reinforce the previously held notion ( e.g. , mcwilliam et al .
2008 ) that the decline in [ o / mg ] with increasing metallicity is likely the result of metallicity dependent yields from massive stars . while we find that the bulge and thick disk exhibit nearly identical [ @xmath3/fe ] ratios at [ fe / h]@xmath130.5 , the bulge stars appear to remain enhanced in [ @xmath3/fe ] by up to 0.10.2 dex higher in [ fe / h ] than the local thick disk .
the bulge [ @xmath3/fe ] ratios at [ fe / h]@xmath20 are well matched to the local thin disk trends .
these results are in agreement with recent differential abundance analyses of microlensed bulge dwarfs ( bensby et al . 2013 ) , and suggest the bulge experienced faster enrichment than the local thick disk . however , similar differential analyses comparing bulge and thick disk giants find no significant differences between the two populations ( melndez et al .
2008 ; alves brito et al . 2010 ; gonzalez et al .
2011 ) . combining the new data set of [ @xmath3/fe ] abundances with those available in the literature
now totals several hundred stars .
however , the combined data set does not reveal any significant population with
anomalous " chemistry , such as the low [ @xmath3/fe ] ratios reminiscent of many present day dwarf galaxy stars .
therefore , we can effectively rule out these types of objects as major contributors to any portion of the present day galactic bulge field population .
this further supports the idea that the galactic bulge is not a merger built system .
similarly , the [ @xmath3/fe ] ratios of the ngc 6553 stars are identical to those of similar metallicity field stars .
this suggests ngc 6553 formed _ in situ _ with the bulge and is not a captured system . with regard to the light , odd
z elements , we find that na and al exhibit discrepant trends as a function of metallicity . in particular , bulge stars exhibit a steady increase in [ na / fe ] with increasing [ fe / h ] , but the [ al / fe ] trend almost exactly matches that of the @xmath3elements ( except oxygen ) . while we do not find any significant field to
field variations in either [ na / fe ] or [ al / fe ] , our results indicate that the bulge and thick disk have different [ na / fe ] abundances at [ fe / h]@xmath130.5 but similar [ al / fe ] .
interestingly , the @xmath3like " behavior of [ al / fe ] contrasts with several previous bulge studies that found [ al / fe ] was enhanced up to [ fe / h][email protected] . instead , our results are in agreement with the abundance patterns of microlensed bulge dwarfs ( e.g. , bensby et al .
the discrepant behavior of na and al suggests metallicity dependent yields from massive stars , and perhaps intermediate mass stars , leads to significantly more production of na than al at high metallicity .
we also find that the ngc 6553 stars have nearly identical [ al / fe ] ratios as similar metallicity field stars , but both the average [ na / fe ] abundance and star to star dispersion of cluster stars are higher .
this suggests ngc 6553 experienced some light element self enrichment , which is typical for globular clusters .
the abundance trends of the fe
peak elements are distinctly different : ( 1 ) the average [ cr / fe ] ratio is essentially solar over the full range in [ fe / h ] and shows no variations over the metallicity range probed here , ( 2 ) both [ co / fe ] and [ ni / fe ] are enhanced by @[email protected] dex at nearly all [ fe / h ] and exhibit some low amplitude , metallicity dependent variations , and ( 3 ) [ cu / fe ] exhibits a large increase from the metal poor to metal rich end of the distribution . in a similar fashion to [ na / fe ] ,
the strong secondary ( metallicity dependent ) production of cu is evident in bulge stars , and the correlation between [ cu / o ] and [ fe / h ] suggests massive stars produce significant portions of cu .
however , cu production from another source ( e.g. , type ia sne ) seems required to explain the high [ cu / fe ] abundances at super solar metallicities . interestingly , at [ fe / h]@xmath22 the [ cr / fe ] trend is identical between the bulge , thick disk , and thin disk , but the heavier fe peak [ x / fe ] ratios appear to all be enhanced in the bulge relative to the local disk
. additionally , the ngc 6553 fe
peak abundance trends are in agreement with similar metallicity field stars . despite predicting [ na / fe ] , [ al / fe ] , and [ ni / fe ] ratios that are too high ,
our adopted baseline bulge chemical enrichment model from kobayashi et al .
( 2006 ; 2011 ) does a reasonable job fitting the abundance trends of the @xmath3 and other fe peak elements .
however , better agreement between the data and model is found when a significant fraction of hne , which produce more fe , are included .
unfortunately , setting the hn fraction is best constrained using abundance patterns at [ fe / h]@xmath131 , where the bulge data are sparse . while a kroupa ( 2008 ) imf provides a reasonable fit to the observed abundance trends , a top heavy imf including strong outflow can not be ruled out .
in contrast , the fe peak abundance data strongly rule out imfs that are truncated to exclude the contributions of stars @xmath440 m@xmath5 , steep imfs ( e.g. , x=1.6 ) , and top heavy imfs that do not include outflow .
we conclude that the bulge likely does not require a particularly unusual imf to explain its present
day abundance patterns , and that its enhanced abundances for several @xmath3 and fe peak elements match model predictions in which the bulge experienced a different enrichment history than the local disk .
we thank the anonymous referee for a careful reading of the manuscript and helpful comments that lead to improvement of the manuscript .
this research has made use of nasa s astrophysics data system bibliographic services .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
cij gratefully acknowledges support from the clay fellowship , administered by the smithsonian astrophysical observatory .
rmr acknowledges support from nsf - ast-1212095 .
ak acknowledges the deutsche forschungsgemeinschaft for funding from emmy noether grant ko 4161/1 .
hinkle , k. , wallace , l. , valenti , j. , & harmer , d. 2000 , visible and near infrared atlas of the arcturus spectrum 3727 - 9300 a ed .
kenneth hinkle , lloyd wallace , jeff valenti , and dianne harmer .
( san francisco : asp ) isbn : 1 - 58381 - 037 - 4 , 2000 + 119799c4 & 18081026@xmath302547354 & 272.042708 & @xmath3025.793167 & & & 12.856 & 11.986 & 11.704 & 4400 & 2.05 & @xmath300.39 & 1.90 & @xmath057.96 & 0.11 + 129499c4 & 18081617@xmath302543192 & 272.067375 & @xmath3025.721972 & & & 12.596 & 11.836 & 11.608 & 4900 & 2.75 & @xmath00.13 & 2.10 & @xmath027.10 & 0.04 + 176772c5 & 18080936@xmath302556168 & 272.039417 & @xmath3025.938028 & & & 11.710 & 10.794 & 10.483 & 4225 & 2.15 & @xmath00.06 & 1.90 & @xmath037.82 & 0.08 + 181349c5 & 18080465@xmath302554356 & 272.019292 & @xmath3025.909917 & & & 12.924 & 12.253 & 11.906 & 4775 & 2.75 & @xmath00.26 & 2.10 & @xmath038.51 & 0.04 + 183783c5 & 18080912@xmath302553420 & 272.037958 & @xmath3025.895028 & & & 12.876 & 12.119 & 11.848 & 4700 & 2.75 & @xmath00.10 & 1.75 & @xmath3097.45 & 0.01 + 184088c5 & 18080726@xmath302553356 & 272.030250 & @xmath3025.893222 & & & 12.789 & 11.869 & 11.640 & 4500 & 2.50 & @xmath300.47 & 1.80 & @xmath0173.45 & 0.22 + 184618c5 & 18080414@xmath302553236 & 272.017208 & @xmath3025.889917 & & & 12.732 & 11.946 & 11.690 & 4600 & 2.15 & @xmath300.59 & 1.55 & @xmath0247.54 & 0.34 + 185169c5 & 18080523@xmath302553111 & 272.021708 & @xmath3025.886389 & & & 12.367 & 11.587 & 11.360 & 4775 & 2.50 & @xmath300.32 & 1.85 & @xmath084.03 & 0.25 + 185357c5 & 18081778@xmath302553068 & 272.074083 & @xmath3025.885250 & & & 12.933 & 12.152 & 11.841 & 4550 & 2.60 & @xmath00.28 & 1.90 & @xmath3060.42 & 0.16 + 185541c5 & 18080978@xmath302553024 & 272.040708 & @xmath3025.884000 & & & 12.865 & 12.106 & 11.865 & 4750 & 3.10 & @xmath00.34 & 1.80 & @xmath3012.63 & 0.02 + 187067c5 & & 272.056958 & @xmath3025.874083 & & & & & & 4850 & 2.30 & @xmath300.95 & 1.35 & @xmath0115.30 & 0.31 + 193190c5 & 18081265@xmath302550046 & 272.052625 & @xmath3025.834556 & & & 12.248 & 11.461 & 11.172 & 4675 & 2.75 & @xmath00.24 & 1.85 & @xmath07.60 & 0.34 + 197366c5 & 18081403@xmath302548290 & 272.058417 & @xmath3025.808028 & & & 12.767 & 11.959 & 11.700 & 4550 & 2.90 & @xmath00.32 & 1.75 & @xmath3073.59 & 0.03 + 215681c6 & 18084446@xmath302557568 & 272.185292 & @xmath3025.965778 & & & 12.938 & 12.160 & 11.944 & 4800 & 2.75 & @xmath00.31 & 1.95 & @xmath054.15 & 0.20 + 216922c6 & 18084514@xmath302557293 & 272.188083 & @xmath3025.958139 & & & 12.598 & 11.652 & 11.472 & 4400 & 2.10 & @xmath300.20 & 2.20 & @xmath30135.08 & 0.16 + 218198c6 & 18082120@xmath302557000 & 272.088333 & @xmath3025.950028 & & & 12.506 & 11.579 & 11.271 & 4225 & 1.95 & @xmath300.32 & 2.10 & @xmath30158.44 & 0.02 + 219909c6 & 18083262@xmath302556215 & 272.135917 & @xmath3025.939306 & & & 12.618 & 11.745 & 11.520 & 4650 & 2.25 & @xmath300.73 & 1.75 & @xmath3062.33 & 0.02 + 221537c6 & 18085002@xmath302555460 & 272.208375 & @xmath3025.929417 & & & 12.881 & 12.057 & 11.835 & 4750 & 2.75 & @xmath00.29 & 1.60 & @xmath3060.86 & 0.04 + 223113c6 & 18084072@xmath302555101 & 272.169667 & @xmath3025.919472 & & & 12.833 & 12.074 & 11.845 & 4800 & 2.55 & @xmath300.15 & 1.75 & @xmath3025.52 & 0.32 + 223343c6 & 18082127@xmath302555048 & 272.088625 & @xmath3025.918028 & & & 12.834 & 12.008 & 11.774 & 4525 & 2.65 & @xmath00.15 & 1.45 & @xmath061.02 & 0.44 + 223621c6 & 18085263@xmath302554587 & 272.219208 & @xmath3025.916306 & & & 12.677 & 11.938 & 11.749 & 5150 & 3.75 & @xmath00.31 & 1.35 & @xmath088.54 & 0.18 + 223722c6 & & 272.183958 & @xmath3025.915722 & & & & & & 4525 & 1.35 & @xmath300.75 & 1.75 & @xmath30114.88 & 0.18 + 224206c6 & 18084690@xmath302554462 & 272.195417 & @xmath3025.912833 & & & 12.857 & 12.069 & 11.853 & 4800 & 3.00 & @xmath00.46 & 1.65 & @xmath30142.44 & 0.13 + 224866c6 & 18083392@xmath302554311 & 272.141417 & @xmath3025.908667 & & & 12.489 & 11.604 & 11.333 & 4350 & 2.10 & @xmath300.22 & 1.75 & @xmath3096.89 & 0.01 + 224951c6 & 18085128@xmath302554294 & 272.213667 & @xmath3025.908167 & & & 12.660 & 11.877 & 11.583 & 4650 & 2.75 & @xmath00.17 & 1.80 & @xmath3087.33 & 0.32 + 225531c6 & 18082878@xmath302554167 & 272.119958 & @xmath3025.904639 & & & 12.613 & 11.670 & 11.452 & 4425 & 1.75 & @xmath300.73 & 1.80 & @xmath0135.60 & 0.20 + 226450c6 & 18083464@xmath302553563 & 272.144333 & @xmath3025.898972 & & & 12.918 & 12.061 & 11.862 & 4500 & 2.50 & @xmath00.27 & 1.80 & @xmath3022.47 & 0.13 + 226850c6 & 18082340@xmath302553475 & 272.097542 & @xmath3025.896500 & & & 12.548 & 11.753 & 11.506 & 4800 & 2.15 & @xmath300.85 & 1.85 & @xmath021.21 & 0.03 + 227867c6 & 18084160@xmath302553245 & 272.173333 & @xmath3025.890167 & & & 12.699 & 11.894 & 11.634 & 4850 & 3.15 & @xmath00.09 & 1.60 & @xmath062.08 & 0.14 + 228466c6 & 18082917@xmath302553116 & 272.121542 & @xmath3025.886583 & & & 12.665 & 11.762 & 11.548 & 4700 & 1.55 & @xmath300.98 & 2.15 & @xmath024.42 & 0.12 + 229507c6 & 18085530@xmath302552488 & 272.230375 & @xmath3025.880222 & & & 12.490 & 11.622 & 11.408 & 4650 & 1.75 & @xmath300.29 & 1.60 & @xmath071.66 & 0.13 + 230424c6 & 18085024@xmath302552281 & 272.209292 & @xmath3025.874500 & & & 12.319 & 11.344 & 11.098 & 4400 & 2.10 & @xmath300.84 & 1.75 & @xmath300.90 & 0.06 + 230483c6 & 18082585@xmath302552269 & 272.107708 & @xmath3025.874139 & & & 12.643 & 11.725 & 11.465 & 4450 & 2.10 & @xmath300.65 & 1.60 & @xmath3019.64 & 0.09 + 231379c6 & 18082290@xmath302552064 & 272.095417 & @xmath3025.868444 & & & 12.787 & 11.906 & 11.675 & 4500 & 2.30 & @xmath300.30 & 1.80 & @xmath036.71 & 0.36 + 231618c6 & 18084424@xmath302552014 & 272.184250 & @xmath3025.867028 & & & 12.540 & 11.776 & 12.711 & 4550 & 2.60 & @xmath00.04 & 1.65 & @xmath022.46 & 0.10 + 232493c6 & & 272.090500 & @xmath3025.861694 & & & & & & 4625 & 2.00 & @xmath301.14 & 1.20 & @xmath0160.24 & 0.05 + 233121c6 & 18083049@xmath302551279 & 272.127000 & @xmath3025.857722 & & & 12.614 & 11.684 & 11.468 & 4500 & 1.90 & @xmath300.50 & 1.90 & @xmath3021.69 & 0.12 + 233560c6 & 18082099@xmath302551174 & 272.087417 & @xmath3025.854833 & & & 12.900 & 11.970 & 11.747 & 4375 & 2.10 & @xmath300.21 & 1.85 & @xmath0170.65 & 0.27 + 233708c6 & 18084123@xmath302551146 & 272.171750 & @xmath3025.853972 & & & 12.393 & 11.437 & 11.064 & 4500 & 2.50 & @xmath00.14 & 1.70 & @xmath057.52 & 0.07 + 240059c6 & 18083133@xmath302548449 & 272.130542 & @xmath3025.812472 & & & 12.410 & 11.452 & 11.258 & 4400 & 1.65 & @xmath300.69 & 1.60 & @xmath30124.52 & 0.52 + 240083c6 & 18082459@xmath302548444 & 272.102500 & @xmath3025.812306 & & & 12.255 & 11.395 & 11.130 & 4425 & 1.65 & @xmath300.58 & 1.70 & @xmath3033.61 & 0.10 + 259050c7 & 18091376@xmath302557523 & 272.307250 & @xmath3025.964528 & & & 12.231 & 11.277 & 11.012 & 4275 & 1.55 & @xmath300.65 & 1.65 & @xmath30124.20 & 0.16 + 259377c7 & 18085927@xmath302557448 & 272.246958 & @xmath3025.962472 & & & 12.735 & 11.939 & 11.723 & 5000 & 3.25 & @xmath00.29 & 1.85 & @xmath025.50 & 0.18 + 260308c7 & 18090858@xmath302557243 & 272.285792 & @xmath3025.956750 & & & 12.659 & 11.825 & 11.611 & 4600 & 2.50 & @xmath00.16 & 1.90 & @xmath3043.66 & 0.10 + 262018c7 & 18091405@xmath302556473 & 272.308583 & @xmath3025.946472 & & & 12.937 & 12.109 & 11.925 & 4700 & 2.75 & @xmath00.08 & 2.00 & @xmath3016.13 & 0.31 + 266442c7 & 18090118@xmath302555154 & 272.254917 & @xmath3025.920917 & & & 12.855 & 11.908 & 11.678 & 4350 & 2.00 & @xmath300.42 & 1.60 & @xmath3062.73 & 0.02 + 270316c7 & 18091628@xmath302553550 & 272.317917 & @xmath3025.898667 & & & 12.675 & 11.766 & 11.538 & 4400 & 1.90 & @xmath300.37 & 1.65 & @xmath301.74 & 0.47 + 275181c7 & 18091022@xmath302552134 & 272.292583 & @xmath3025.870389 & & & 12.478 & 11.578 & 11.366 & 4375 & 1.70 & @xmath300.59 & 1.75 & @xmath0253.05 & 0.05 + 277490c7 & 18090629@xmath302551235 & 272.276208 & @xmath3025.856500 & & & 12.814 & 11.861 & 11.606 & 4250 & 1.55 & @xmath300.56 & 1.65 & @xmath076.93 & 0.38 + 278419c7 & 18085959@xmath302551037 & 272.248292 & @xmath3025.850972 & & & 12.777 & 12.019 & 11.714 & 4725 & 2.75 & @xmath00.20 & 1.55 & @xmath3068.67 & 0.05 + 282804c7 & 18091653@xmath302549261 & 272.318917 & @xmath3025.823889 & & & 12.468 & 11.614 & 11.351 & 4575 & 2.50 & @xmath00.10 & 1.70 & @xmath0135.83 & 0.04 + 286252c7 & 18090052@xmath302548070 & 272.252167 & @xmath3025.801861 & & & 12.579 & 11.777 & 11.563 & 4700 & 3.10 & @xmath00.43 & 1.40 & @xmath3076.01 & 0.06 + 45512c2 & 18090473@xmath302545129 & 272.269792 & @xmath3025.753556 & & & 12.623 & 11.821 & 11.552 & 4500 & 2.15 & @xmath300.27 & 1.75 & @xmath096.20 & 0.03 + 47188c2 & 18090688@xmath302544307 & 272.278667 & @xmath3025.741833 & & & 12.258 & 11.324 & 11.057 & 4350 & 1.70 & @xmath300.52 & 1.85 & @xmath30128.90 & 0.34 + 77186c3 & 18082835@xmath302547382 & 272.118125 & @xmath3025.793889 & & & 12.942 & 12.153 & 11.890 & 4750 & 3.15 & @xmath00.41 & 1.85 & @xmath305.75 & 0.10 + 77707c3 & 18083637@xmath302547256 & 272.151542 & @xmath3025.790306 & & & 12.479 & 11.623 & 11.448 & 4825 & 3.15 & @xmath00.44 & 1.65 & @xmath04.81 & 0.14 + 80582c3 & 18085189@xmath302546213 & 272.216208 & @xmath3025.772556 & & & 12.765 & 11.945 & 11.686 & 5100 & 3.55 & @xmath00.07 & 1.60 & @xmath3039.87 & 0.37 + 81644c3 & 18085454@xmath302545569 & 272.227208 & @xmath3025.765806 & & & 12.631 & 11.753 & 11.430 & 4600 & 2.85 & @xmath00.21 & 1.70 & @xmath052.17 & 0.11 + 82227c3 & 18084141@xmath302545439 & 272.172500 & @xmath3025.762139 & & & 12.676 & 11.828 & 11.642 & 4900 & 3.25 & @xmath00.25 & 1.65 & @xmath3014.02 & 0.18 + 83531c3 & 18084381@xmath302545145 & 272.182500 & @xmath3025.754000 & & & 12.535 & 11.692 & 11.431 & 4725 & 3.25 & @xmath00.45 & 1.80 & @xmath3026.11 & 0.22 + 84255c3 & 18082150@xmath302544574 & 272.089583 & @xmath3025.749278 & & & 12.583 & 11.668 & 11.423 & 4350 & 1.50 & @xmath300.68 & 1.45 & @xmath30221.43 & 0.03 + 86757c3 & 18083011@xmath302544007 & 272.125458 & @xmath3025.733500 & & & 12.467 & 11.541 & 11.300 & 4400 & 2.20 & @xmath300.02 & 1.90 & @xmath072.12 & 0.01 + 88522c3 & 18083582@xmath302543214 & 272.149250 & @xmath3025.722611 & & & 12.463 & 11.613 & 11.383 & 4700 & 3.00 & @xmath00.39 & 1.90 & @xmath305.87 & 0.23 + + 225847c6 & 18085574@xmath302554094 & 272.232292 & @xmath3025.902583 & & & 12.895 & 12.000 & 11.757 & 4500 & 2.25 & @xmath300.11 & 1.45 & @xmath303.26 & 0.24 + 227379c6 & 18082809@xmath302553354 & 272.117083 & @xmath3025.893167 & & & 12.947 & 12.011 & 11.779 & 4350 & 2.25 & @xmath300.05 & 1.55 & @xmath302.50 & 0.13 + 228407c6 & 18085354@xmath302553128 & 272.223042 & @xmath3025.886889 & & & 12.718 & 11.781 & 11.547 & 4415 & 1.95 & @xmath300.22 & 1.75 & @xmath00.21 & 0.16 + 230208c6 & 18085430@xmath302552332 & 272.226208 & @xmath3025.875917 & & & 12.846 & 12.020 & 11.842 & 4900 & 3.05 & @xmath00.00 & 1.40 & @xmath01.43 & 0.46 + 239284c6 & 18083226@xmath302549034 & 272.134417 & @xmath3025.817611 & & & 12.438 & 11.512 & 11.232 & 4300 & 2.00 & @xmath300.14 & 1.80 & @xmath309.48 & 0.43 + 265795c7 & 18092018@xmath302555283 & 272.334125 & @xmath3025.924556 & & & 12.757 & 11.959 & 11.788 & 4600 & 2.50 & @xmath300.07 & 1.55 & @xmath3012.45 & 0.03 + 268360c7 & 18091379@xmath302554357 & 272.308000 & @xmath3025.909833 & & & 12.075 & 11.278 & 10.987 & 4750 & 3.15 & @xmath300.08 & 1.35 & @xmath03.63 & 0.71 + 268493c7 & 18090837@xmath302554327 & 272.284917 & @xmath3025.909056 & & & 12.647 & 11.734 & 11.552 & 4500 & 2.30 & @xmath300.20 & 1.90 & @xmath301.86 & 0.11 + 271021c7 & 18090661@xmath302553404 & 272.277667 & @xmath3025.894472 & & & 12.544 & 11.775 & 11.591 & 4750 & 2.50 & @xmath300.06 & 1.65 & @xmath00.22 & 0.06 + 271400c7 & 18091320@xmath302553328 & 272.305083 & @xmath3025.892389 & & & 12.418 & 11.519 & 11.325 & 4525 & 2.50 & @xmath300.19 & 1.90 & @xmath304.03 & 0.22 + 77182c3 & 18084966@xmath302547388 & 272.206875 & @xmath3025.794056 & & & 12.806 & 11.927 & 11.670 & 4700 & 2.50 & @xmath300.09 & 1.75 & @xmath03.44 & 0.17 + 85597c3 & 18082369@xmath302544268 & 272.098708 & @xmath3025.740778 & & & 12.723 & 11.896 & 11.655 & 4400 & 2.25 & @xmath300.15 & 1.55 & @xmath00.24 & 0.17 + + 1156c2 & 18355601@xmath303433364 & 278.983417 & @xmath3034.560111 & 15.93 & 1.534 & 13.092 & 12.380 & 12.280 & 4300 & 1.80 & @xmath300.46 & 1.45 & @xmath072.84 & 0.01 + 1407c3 & 18345053@xmath303433226 & 278.710542 & @xmath3034.556306 & 15.76 & 1.155 & 13.490 & 13.008 & 12.995 & 5125 & 2.40 & @xmath300.62 & 1.45 & @xmath30141.73 & 0.09 + 1491c7 & 18353352@xmath303449238 & 278.889708 & @xmath3034.823306 & 15.80 & 1.419 & 13.184 & 12.562 & 12.461 & 4700 & 2.25 & @xmath300.32 & 1.85 & @xmath3031.46 & 0.62 + 1554c7 & 18354037@xmath303449160 & 278.918208 & @xmath3034.821139 & 15.11 & 1.166 & 12.887 & 12.390 & 12.333 & 5050 & 2.20 & @xmath300.67 & 1.50 & @xmath07.77 & 0.18 + 166c3 & 18344539@xmath303436070 & 278.689125 & @xmath3034.602000 & 15.90 & 1.267 & 13.466 & 12.945 & 12.873 & 4850 & 2.20 & @xmath300.83 & 1.30 & @xmath054.10 & 0.12 + 1754c3 & 18344892@xmath303432327 & 278.703833 & @xmath3034.542417 & 15.43 & 1.301 & 12.955 & 12.409 & 12.298 & 4900 & 2.50 & @xmath300.25 & 1.65 & @xmath018.44 & 0.06 + 1814c1 & 18361156@xmath303432060 & 279.048250 & @xmath3034.535000 & 15.85 & 1.427 & 13.229 & 12.567 & 12.421 & 4650 & 2.60 & @xmath300.30 & 1.50 & @xmath057.13 & 0.01 + 1876c2 & 18354937@xmath303431562 & 278.955750 & @xmath3034.532250 & 15.86 & 1.106 & 13.679 & 13.187 & 13.147 & 5000 & 2.50 & @xmath301.03 & 1.10 & @xmath0202.57 & 0.03 + 1917c1 & 18361248@xmath303431515 & 279.052083 & @xmath3034.530972 & 16.22 & 1.426 & 13.562 & 12.975 & 12.850 & 4675 & 2.70 & @xmath00.18 & 1.50 & @xmath3053.87 & 0.01 + 1918c1 & 18360752@xmath303431511 & 279.031375 & @xmath3034.530833 & 15.83 & 1.248 & 13.428 & 12.875 & 12.791 & 4900 & 2.35 & @xmath300.44 & 1.60 & @xmath042.10 & 0.05 + 201583c3 & 18345139@xmath303431404 & 278.714125 & @xmath3034.527889 & 15.79 & 1.258 & 13.411 & 12.848 & 12.813 & 4800 & 1.90 & @xmath301.11 & 0.90 & @xmath059.07 & 0.08 + 2110c7 & 18353609@xmath303447564 & 278.900417 & @xmath3034.799056 & 15.59 & 1.405 & 13.013 & 12.367 & 12.338 & 4675 & 3.00 & @xmath00.02 & 1.30 & @xmath306.53 & 0.10 + 2178c7 & 18354156@xmath303447486 & 278.923167 & @xmath3034.796889 & 15.77 & 1.371 & 13.195 & 12.619 & 12.505 & 4700 & 2.80 & @xmath00.00 & 1.80 & @xmath3012.34 & 0.10 + 2200c3 & 18345778@xmath303431359 & 278.740708 & @xmath3034.526639 & 16.15 & 1.315 & 13.588 & 13.073 & 13.014 & 4975 & 2.70 & @xmath300.10 & 1.80 & @xmath306.43 & 0.11 + 2220c7 & 18353072@xmath303447426 & 278.878000 & @xmath3034.795194 & 15.62 & 1.233 & 13.230 & 12.739 & 12.685 & 5150 & 2.75 & @xmath300.17 & 1.60 & @xmath3081.33 & 0.06 + 222c3 & 18351488@xmath303436002 & 278.812083 & @xmath3034.600111 & 15.50 & 1.295 & 12.984 & 12.449 & 12.329 & 4950 & 2.60 & @xmath300.43 & 1.90 & @xmath090.22 & 0.53 + 2335c2 & 18355884@xmath303430461 & 278.995208 & @xmath3034.512778 & 15.81 & 1.270 & 13.385 & 12.778 & 12.718 & 4750 & 2.05 & @xmath300.90 & 1.50 & @xmath30119.07 & 0.12 + 2407c2 & 18360198@xmath303430369 & 279.008292 & @xmath3034.510222 & 15.49 & 1.173 & 13.232 & 12.653 & 12.605 & 4975 & 2.30 & @xmath300.64 & 1.70 & @xmath30145.30 & 0.15 + 2422c7 & 18354805@xmath303447154 & 278.950250 & @xmath3034.787667 & 15.89 & 1.481 & 13.176 & 12.551 & 12.476 & 4400 & 2.15 & @xmath00.17 & 1.90 & @xmath3021.36 & 0.07 + 2470c3 & 18352151@xmath303431018 & 278.839667 & @xmath3034.517167 & 16.25 & 1.375 & 13.661 & 13.048 & 12.952 & 4800 & 3.10 & @xmath00.00 & 1.35 & @xmath089.04 & 0.18 + 2502c3 & 18345442@xmath303430566 & 278.726750 & @xmath3034.515694 & 15.97 & 1.330 & 13.424 & 12.833 & 12.762 & 4750 & 2.25 & @xmath300.62 & 1.40 & @xmath0255.23 & 0.44 + 2532c6 & 18351318@xmath303446482 & 278.804917 & @xmath3034.780139 & 15.71 & 1.345 & 13.123 & 12.610 & 12.467 & 4750 & 2.50 & @xmath300.17 & 1.90 & @xmath3042.27 & 0.16 + 2580c6 & 18351524@xmath303446412 & 278.813500 & @xmath3034.778167 & 16.06 & 1.566 & 13.184 & 12.595 & 12.406 & 4575 & 2.85 & @xmath00.30 & 2.00 & @xmath3016.54 & 0.24 + 2769c3 & 18345752@xmath303430213 & 278.739667 & @xmath3034.505917 & 15.78 & 1.295 & 13.311 & 12.797 & 12.657 & 4900 & 3.15 & @xmath00.06 & 1.25 & @xmath3051.43 & 0.48 + 2772c7 & 18354807@xmath303446257 & 278.950292 & @xmath3034.773833 & 15.10 & 1.258 & 12.692 & 12.137 & 12.072 & 4900 & 2.75 & @xmath300.25 & 1.65 & @xmath3079.69 & 0.09 + 2812c8 & 18360927@xmath303446233 & 279.038667 & @xmath3034.773194 & 15.63 & 1.458 & 12.948 & 12.274 & 12.101 & 4475 & 2.00 & @xmath300.65 & 1.55 & @xmath090.18 & 0.36 + 2947c3 & 18350059@xmath303429557 & 278.752458 & @xmath3034.498778 & 16.02 & 1.204 & 13.660 & 13.157 & 13.120 & 5150 & 2.60 & @xmath300.50 & 1.95 & @xmath3098.82 & 0.15 + 2948c7 & 18352452@xmath303446010 & 278.852208 & @xmath3034.767000 & 15.53 & 1.679 & 12.485 & 11.838 & 11.684 & 4275 & 2.00 & @xmath00.12 & 1.65 & @xmath3022.63 & 0.13 + 3018c3 & 18350358@xmath303429453 & 278.764958 & @xmath3034.495889 & 15.49 & 1.404 & 12.836 & 12.190 & 12.033 & 4650 & 2.40 & @xmath300.54 & 1.45 & @xmath3064.70 & 0.13 + 3035c7 & 18355135@xmath303445453 & 278.964000 & @xmath3034.762639 & 15.68 & 1.191 & 13.412 & 12.868 & 12.808 & 5050 & 2.50 & @xmath300.44 & 1.45 & @xmath08.23 & 0.07 + 3091c8 & 18360886@xmath303445437 & 279.036958 & @xmath3034.762167 & 15.32 & 1.236 & 12.938 & 12.390 & 12.306 & 4900 & 3.00 & @xmath300.50 & 1.45 & @xmath076.91 & 0.13 + 3101c7 & 18353713@xmath303445344 & 278.904708 & @xmath3034.759611 & 15.86 & 1.233 & 13.523 & 12.939 & 12.880 & 4950 & 2.65 & @xmath300.40 & 1.70 & @xmath067.03 & 0.02 + 3142c3 & 18350948@xmath303429282 & 278.789500 & @xmath3034.491139 & 15.95 & 1.287 & 13.481 & 12.949 & 12.831 & 4900 & 2.65 & @xmath300.22 & 1.75 & @xmath3018.58 & 0.02 + 3161c3 & 18351942@xmath303429266 & 278.830917 & @xmath3034.490722 & 15.46 & 1.306 & 12.968 & 12.367 & 12.313 & 5100 & 3.20 & @xmath300.15 & 1.70 & @xmath3029.96 & 0.08 + 3191c7 & 18355560@xmath303445217 & 278.981708 & @xmath3034.756083 & 15.40 & 1.243 & 13.051 & 12.479 & 12.431 & 4950 & 2.50 & @xmath300.37 & 1.60 & @xmath064.72 & 0.09 + 3201c6 & 18350301@xmath303445053 & 278.762583 & @xmath3034.751528 & 15.47 & 1.211 & 13.078 & 12.624 & 12.467 & 5200 & 3.50 & @xmath00.04 & 1.35 & @xmath3071.88 & 0.11 + 3238c6 & 18351698@xmath303444595 & 278.820792 & @xmath3034.749889 & 15.73 & 1.322 & 13.174 & 12.632 & 12.523 & 4900 & 3.00 & @xmath300.22 & 1.35 & @xmath3065.42 & 0.19 + 3267c3 & 18352206@xmath303429112 & 278.841917 & @xmath3034.486444 & 15.07 & 1.349 & 12.525 & 11.923 & 11.806 & 4700 & 2.65 & @xmath00.04 & 1.40 & @xmath30119.57 & 0.22 + 3515c3 & 18345072@xmath303428380 & 278.711000 & @xmath3034.477200 & & & 12.658 & 12.099 & 11.965 & 4750 & 2.70 & @xmath00.01 & 1.80 & @xmath3014.32 & 0.08 + 3558c6 & 18351293@xmath303444170 & 278.803875 & @xmath3034.738139 & 16.03 & 1.341 & 13.526 & 12.901 & 12.780 & 4800 & 2.75 & @xmath300.15 & 1.75 & @xmath021.61 & 0.25 + 3690c7 & 18352834@xmath303444085 & 278.868125 & @xmath3034.735750 & 14.97 & 1.208 & 12.606 & 12.088 & 12.009 & 4800 & 1.45 & @xmath301.46 & 1.50 & @xmath040.71 & 0.07 + 3711c7 & 18354105@xmath303444059 & 278.921042 & @xmath3034.735028 & 15.71 & 1.097 & 13.526 & 13.038 & 13.004 & 5350 & 3.70 & @xmath300.49 & 1.25 & @xmath03.48 & 0.21 + 3733c3 & 18351799@xmath303428093 & 278.824958 & @xmath3034.469250 & 15.10 & 1.423 & 12.436 & 11.795 & 11.722 & 4750 & 2.60 & @xmath300.18 & 1.70 & @xmath3045.51 & 0.01 + 3796c6 & 18345825@xmath303443440 & 278.742708 & @xmath3034.728944 & 15.53 & 1.454 & 12.839 & 12.133 & 12.041 & 4500 & 1.85 & @xmath300.82 & 1.65 & @xmath3038.50 & 0.21 + 3965c6 & 18345917@xmath303443219 & 278.746583 & @xmath3034.722806 & 15.46 & 1.663 & 12.442 & 11.705 & 11.570 & 4400 & 2.35 & @xmath300.08 & 1.70 & @xmath3097.46 & 0.27 + 4085c3 & 18351357@xmath303427240 & 278.806542 & @xmath3034.456667 & 16.13 & 1.513 & 13.330 & 12.675 & 12.564 & 4550 & 1.90 & @xmath300.58 & 1.65 & @xmath307.91 & 0.12 + 4217c6 & 18345512@xmath303442462 & 278.729667 & @xmath3034.712861 & 15.78 & 1.461 & 13.065 & 12.390 & 12.213 & 4500 & 2.15 & @xmath300.57 & 1.70 & @xmath3025.21 & 0.08 + 4263c6 & 18350180@xmath303442404 & 278.757542 & @xmath3034.711278 & 15.87 & 1.329 & 13.333 & 12.756 & 12.650 & 5000 & 3.05 & @xmath300.34 & 1.90 & @xmath3090.50 & 0.07 + 431c2 & 18355359@xmath303435209 & 278.973333 & @xmath3034.589194 & 15.63 & 1.175 & 13.302 & 12.805 & 12.744 & 5100 & 3.85 & @xmath300.03 & 1.20 & @xmath032.21 & 0.07 + 4365c3 & 18352116@xmath303426510 & 278.838208 & @xmath3034.447500 & 16.08 & 1.148 & 13.854 & 13.339 & 13.339 & 4950 & 1.75 & @xmath300.80 & 1.65 & @xmath0202.51 & 0.25 + 4478c8 & 18361669@xmath303442257 & 279.069583 & @xmath3034.707167 & 15.79 & 1.260 & 13.441 & 12.906 & 12.751 & 4900 & 2.25 & @xmath300.36 & 1.65 & @xmath061.90 & 0.17 + 455c1 & 18361784@xmath303435209 & 279.074375 & @xmath3034.589194 & 15.67 & 1.174 & 13.413 & 12.878 & 12.798 & 5000 & 2.70 & @xmath300.61 & 1.50 & @xmath30168.72 & 0.05 + 4612c6 & 18350115@xmath303441540 & 278.754792 & @xmath3034.698361 & 16.02 & 1.264 & 13.580 & 12.968 & 12.908 & 4950 & 2.60 & @xmath300.78 & 1.30 & @xmath035.76 & 0.21 + 4740c8 & 18361928@xmath303441470 & 279.080417 & @xmath3034.696417 & 15.80 & 1.208 & 13.515 & 13.011 & 12.826 & 5000 & 2.30 & @xmath300.49 & 1.60 & @xmath0181.98 & 0.17 + 4876c6 & 18351228@xmath303441180 & 278.801208 & @xmath3034.688361 & 16.08 & 1.303 & 13.583 & 12.978 & 12.887 & 4800 & 2.70 & @xmath300.59 & 1.35 & @xmath08.46 & 0.01 + 5319c6 & 18344324@xmath303440187 & 278.680167 & @xmath3034.671889 & 15.58 & 1.447 & 12.877 & 12.249 & 12.110 & 4600 & 2.45 & @xmath300.40 & 1.50 & @xmath30280.09 & 0.22 + 5351c8 & 18360799@xmath303440158 & 279.033333 & @xmath3034.671111 & 15.83 & 1.382 & 13.355 & 12.641 & 12.494 & 4700 & 2.65 & @xmath300.08 & 1.55 & @xmath3079.63 & 0.24 + 5400c8 & 18361236@xmath303440076 & 279.051542 & @xmath3034.668806 & 15.96 & 1.113 & 13.820 & 13.357 & 13.236 & 5050 & 3.75 & @xmath00.12 & 1.30 & @xmath3029.93 & 0.10 + 5487c8 & 18361461@xmath303439566 & 279.060958 & @xmath3034.665750 & 16.16 & 1.339 & 13.645 & 13.067 & 12.971 & 4750 & 2.30 & @xmath300.47 & 1.75 & @xmath045.61 & 0.01 + 5543c6 & 18351584@xmath303439495 & 278.816042 & @xmath3034.663778 & 16.07 & 1.459 & 13.311 & 12.694 & 12.519 & 4400 & 1.60 & @xmath300.17 & 2.00 & @xmath306.45 & 0.04 + 5588c6 & 18350642@xmath303439420 & 278.776750 & @xmath3034.661694 & 15.87 & 1.282 & 13.371 & 12.812 & 12.718 & 4900 & 2.60 & @xmath300.20 & 1.60 & @xmath054.21 & 0.06 + 5664c6 & 18345260@xmath303439317 & 278.719167 & @xmath3034.658833 & 15.43 & 1.330 & 12.862 & 12.341 & 12.241 & 4800 & 2.75 & @xmath00.23 & 1.60 & @xmath3082.41 & 0.01 + 5908c6 & 18344801@xmath303438540 & 278.700000 & @xmath3034.648333 & 15.66 & 1.151 & 13.444 & 12.982 & 12.844 & 5150 & 2.50 & @xmath300.67 & 1.60 & @xmath30138.69 & 0.05 + 5977c6 & 18351222@xmath303438453 & 278.800917 & @xmath3034.645917 & 15.31 & 1.334 & 12.714 & 12.125 & 12.026 & 4650 & 2.45 & @xmath00.19 & 1.65 & @xmath307.78 & 0.16 + 5980c6 & 18344397@xmath303438441 & 278.683250 & @xmath3034.645611 & 15.93 & 1.242 & 13.459 & 12.912 & 12.849 & 4950 & 3.45 & @xmath300.17 & 1.25 & @xmath3045.00 & 0.08 + 608c1 & 18361368@xmath303434559 & 279.057042 & @xmath3034.582194 & 15.13 & 1.133 & 12.938 & 12.450 & 12.348 & 5000 & 2.45 & @xmath301.74 & 1.00 & @xmath030.98 & 0.10 + 6090c6 & 18352004@xmath303438289 & 278.833542 & @xmath3034.641333 & 16.15 & 1.479 & 13.361 & 12.735 & 12.575 & 4500 & 2.25 & @xmath300.02 & 1.90 & @xmath042.30 & 0.16 + 6164c6 & 18345884@xmath303438175 & 278.745167 & @xmath3034.638194 & 15.25 & 1.265 & 12.830 & 12.234 & 12.183 & 4900 & 3.25 & @xmath300.13 & 0.95 & @xmath3060.24 & 0.06 + 6230c5 & 18343398@xmath303438536 & 278.641583 & @xmath3034.648222 & 15.06 & 1.154 & 12.814 & 12.371 & 12.284 & 5050 & 3.60 & @xmath00.20 & 1.25 & @xmath3036.65 & 0.10 + 6263c6 & 18351420@xmath303438060 & 278.809208 & @xmath3034.635000 & 16.01 & 1.648 & 12.949 & 12.313 & 12.147 & 4250 & 2.30 & @xmath00.48 & 1.95 & @xmath307.85 & 0.03 + 6391c8 & 18362357@xmath303437379 & 279.098250 & @xmath3034.627167 & 15.28 & 1.140 & 13.105 & 12.554 & 12.534 & 5000 & 2.20 & @xmath300.63 & 1.55 & @xmath3092.33 & 0.09 + 6419c5 & 18343096@xmath303438299 & 278.629000 & @xmath3034.641667 & 15.40 & 1.390 & 12.821 & 12.183 & 12.082 & 4700 & 2.60 & @xmath300.24 & 1.55 & @xmath309.31 & 0.04 + 6426c8 & 18361008@xmath303437330 & 279.042083 & @xmath3034.625833 & 15.79 & 1.280 & 13.374 & 12.825 & 12.667 & 4950 & 2.95 & @xmath300.40 & 1.60 & @xmath030.84 & 0.28 + 6505c6 & 18351867@xmath303437309 & 278.827833 & @xmath3034.625222 & 16.04 & 1.232 & 13.778 & 13.038 & 12.999 & 5000 & 2.70 & @xmath300.38 & 1.50 & @xmath3041.77 & 0.11 + 650c2 & 18355679@xmath303434481 & 278.986667 & @xmath3034.580083 & 15.29 & 1.530 & 12.462 & 11.745 & 11.632 & 4350 & 1.55 & @xmath300.67 & 1.60 & @xmath045.55 & 0.06 + 6549c6 & 18350342@xmath303437243 & 278.764292 & @xmath3034.623417 & 15.70 & 1.320 & 13.270 & 12.614 & 12.506 & 4900 & 3.35 & @xmath00.33 & 1.30 & @xmath3044.42 & 0.03 + 6637c8 & 18360757@xmath303437029 & 279.031583 & @xmath3034.617472 & 16.07 & 1.296 & 13.599 & 13.056 & 12.978 & 4800 & 2.50 & @xmath300.26 & 1.60 & @xmath3032.65 & 0.20 + 6717c6 & 18345238@xmath303437018 & 278.718292 & @xmath3034.617167 & 15.58 & 1.283 & 13.127 & 12.532 & 12.465 & 4900 & 3.30 & @xmath300.19 & 1.20 & @xmath0132.37 & 0.03 + 6828c7 & 18355625@xmath303436545 & 278.984375 & @xmath3034.615111 & 16.09 & 1.360 & 13.467 & 12.896 & 12.808 & 4650 & 2.50 & @xmath00.10 & 1.70 & @xmath02.17 & 0.02 + 6913c7 & 18355267@xmath303436435 & 278.969458 & @xmath3034.612028 & 15.95 & 1.244 & 13.520 & 13.003 & 12.893 & 5000 & 2.75 & @xmath300.27 & 1.70 & @xmath304.76 & 0.07 + 867c3 & 18350909@xmath303434346 & 278.787917 & @xmath3034.576306 & 15.76 & 1.228 & 13.378 & 12.839 & 12.733 & 5000 & 2.60 & @xmath300.47 & 1.60 & @xmath30136.48 & 0.25 & 6300.30 & 0.00 & @xmath309.750 & 8.69 & 8.63 & @xmath00.44 + na i & 6154.23 & 2.10 & @xmath301.560 & 6.33 & 5.89 & @xmath00.06 + na i & 6160.75 & 2.10 & @xmath301.210 & 6.33 & 5.89 & @xmath00.06 + mg i & 6318.71 & 5.10 & @xmath302.010 & 7.58 & 7.38 & @xmath00.30 + mg i & 6319.24 & 5.10 & @xmath302.250 & 7.58 & 7.38 & @xmath00.30 + mg i & 6319.49 & 5.10 & @xmath302.730 & 7.58 & 7.38 & @xmath00.30 + al i & 6696.02 & 3.14 & @xmath301.570 & 6.47 & 6.28 & @xmath00.31 + al i & 6698.67 & 3.14 & @xmath301.890 & 6.47 & 6.28 & @xmath00.31
+ si i & 5645.61 & 4.93 & @xmath302.090 & 7.55 & 7.38 & @xmath00.33
+ si i & 5654.92 & 5.61 & @xmath301.714 & 7.55 & 7.38 & @xmath00.33
+ si i & 5665.56 & 4.92 & @xmath301.910 & 7.55 & 7.38 & @xmath00.33
+ si i & 5666.68 & 5.62 & @xmath301.805 & 7.55 & 7.38 & @xmath00.33
+ si i & 5690.43 & 4.93 & @xmath301.910 & 7.55 & 7.38 & @xmath00.33
+ si i & 5701.10 & 4.93 & @xmath302.080 & 7.55 & 7.38 & @xmath00.33
+ si i & 6142.48 & 5.62 & @xmath301.575 & 7.55 & 7.38 & @xmath00.33
+ si i & 6145.02 & 5.62 & @xmath301.460 & 7.55 & 7.38 & @xmath00.33
+ si i & 6155.13 & 5.62 & @xmath300.774 & 7.55 & 7.38 & @xmath00.33
+ si i & 6155.69 & 5.62 & @xmath302.352 & 7.55 & 7.38 & @xmath00.33
+ si i & 6195.43 & 5.87 & @xmath301.560 & 7.55 & 7.38 & @xmath00.33
+ si i & 6237.32 & 5.61 & @xmath301.115 & 7.55 & 7.38 & @xmath00.33
+ si i & 6244.47 & 5.62 & @xmath301.303 & 7.55 & 7.38 & @xmath00.33
+ si i & 6721.85 & 5.86 & @xmath301.016 & 7.55 & 7.38 & @xmath00.33 + ca i & 5594.46 & 2.52 & @xmath300.370 & 6.36 & 6.07 & @xmath00.21 + ca i & 5601.28 & 2.53 & @xmath300.463 & 6.36 & 6.07 & @xmath00.21 + ca i & 5715.82 & 2.71 & @xmath303.386 & 6.36 & 6.07 & @xmath00.21 + ca i & 6122.22 & 1.89 & @xmath300.466 & 6.36 & 6.07 & @xmath00.21 + ca i & 6156.02 & 2.52 & @xmath302.637 & 6.36 & 6.07 & @xmath00.21 + ca i & 6161.30 & 2.52 & @xmath301.246 & 6.36 & 6.07 & @xmath00.21 + ca i & 6162.17 & 1.90 & @xmath300.210 & 6.36 & 6.07 & @xmath00.21 + ca i & 6166.44 & 2.52 & @xmath301.262 & 6.36 & 6.07 & @xmath00.21 + ca i & 6169.04 & 2.52 & @xmath300.837 & 6.36 & 6.07 & @xmath00.21 + ca i & 6169.56 & 2.53 & @xmath300.628 & 6.36 & 6.07 & @xmath00.21 + cr i & 5628.64 & 3.42 & @xmath300.832 & 5.67 & 5.09 & @xmath300.08 + cr i & 5642.36 & 3.86 & @xmath300.840 & 5.67 & 5.09 & @xmath300.08 + cr i & 5648.26 & 3.83 & @xmath300.980 & 5.67 & 5.09 & @xmath300.08 + cr i & 5674.17 & 3.56 & @xmath301.507 & 5.67 & 5.09 & @xmath300.08 + cr i & 5712.77 & 3.01 & @xmath301.107 & 5.67 & 5.09 & @xmath300.08 + cr i & 5719.82 & 3.01 & @xmath301.660 & 5.67 & 5.09 & @xmath300.08 + cr i & 5729.21 & 3.85 & @xmath301.038 & 5.67 & 5.09 & @xmath300.08 + cr i & 5783.06 & 3.32 & @xmath300.510 & 5.67 & 5.09 & @xmath300.08 + cr i & 5784.97 & 3.32 & @xmath300.440 & 5.67 & 5.09 & @xmath300.08 + cr i & 5787.92 & 3.32 & @xmath300.183 & 5.67 & 5.09 & @xmath300.08 + cr i & 5788.38 & 3.01 & @xmath301.524 & 5.67 & 5.09 & @xmath300.08 + cr i & 5790.65 & 1.00 & @xmath304.033 & 5.67 & 5.09 & @xmath300.08 + cr i & 6330.09 & 0.94 & @xmath303.000 & 5.67 & 5.09 & @xmath300.08 + cr i & 6630.01 & 1.03 & @xmath303.560 & 5.67 & 5.09 & @xmath300.08 + cr i & 6729.73 & 4.39 & @xmath300.753 & 5.67 & 5.09 & @xmath300.08 + fe i & 5595.06 & 5.06 & @xmath301.490 & 7.52 & 7.02 & @xmath300.50 + fe i & 5607.66 & 4.15 & @xmath302.260 & 7.52 & 7.02 & @xmath300.50 + fe i & 5608.97 & 4.21 & @xmath302.240 & 7.52 & 7.02 & @xmath300.50 + fe i & 5611.36 & 3.63 & @xmath303.010 & 7.52 & 7.02 & @xmath300.50 + fe i & 5614.28 & 5.09 & @xmath301.298 & 7.52 & 7.02 & @xmath300.50 + fe i & 5615.30 & 2.59 & @xmath302.268 & 7.52 & 7.02 & @xmath300.50 + fe i & 5615.64 & 3.33 & @xmath300.170 & 7.52 & 7.02 & @xmath300.50 + fe i & 5618.63 & 4.21 & @xmath301.456 & 7.52 & 7.02 & @xmath300.50 + fe i & 5619.22 & 3.69 & @xmath303.170 & 7.52 & 7.02 & @xmath300.50 + fe i & 5619.60 & 4.39 & @xmath301.420 & 7.52 & 7.02 & @xmath300.50 + fe i & 5622.94 & 3.64 & @xmath302.986 & 7.52 & 7.02 & @xmath300.50 + fe i & 5624.02 & 4.39 & @xmath301.230 & 7.52 & 7.02 & @xmath300.50 + fe i & 5624.54 & 3.42 & @xmath300.440 & 7.52 & 7.02 & @xmath300.50 + fe i & 5627.08 & 4.18 & @xmath302.920 & 7.52 & 7.02 & @xmath300.50 + fe i & 5633.95 & 4.99 & @xmath300.310 & 7.52 & 7.02 & @xmath300.50 + fe i & 5635.82 & 4.26 & @xmath301.640 & 7.52 & 7.02 & @xmath300.50 + fe i & 5636.70 & 3.64 & @xmath302.630 & 7.52 & 7.02 & @xmath300.50 + fe i & 5638.26 & 4.22 & @xmath300.820 & 7.52 & 7.02 & @xmath300.50 + fe i & 5641.43 & 4.26 & @xmath300.890 & 7.52 & 7.02 & @xmath300.50 + fe i & 5646.68 & 4.26 & @xmath302.440 & 7.52 & 7.02 & @xmath300.50 + fe i & 5649.99 & 5.10 & @xmath300.770 & 7.52 & 7.02 & @xmath300.50 + fe i & 5650.71 & 5.08 & @xmath300.810 & 7.52 & 7.02 & @xmath300.50 + fe i & 5651.47 & 4.47 & @xmath301.850 & 7.52 & 7.02 & @xmath300.50 + fe i & 5652.01 & 4.22 & @xmath303.010 & 7.52 & 7.02 & @xmath300.50 + fe i & 5652.32 & 4.26 & @xmath301.870 & 7.52 & 7.02 & @xmath300.50 + fe i & 5653.86 & 4.39 & @xmath301.480 & 7.52 & 7.02 & @xmath300.50 + fe i & 5655.18 & 5.06 & @xmath300.600 & 7.52 & 7.02 & @xmath300.50 + fe i & 5661.35 & 4.28 & @xmath301.856 & 7.52 & 7.02 & @xmath300.50 + fe i & 5661.97 & 4.26 & @xmath302.770 & 7.52 & 7.02 & @xmath300.50 + fe i & 5662.52 & 4.18 & @xmath300.563 & 7.52 & 7.02 & @xmath300.50 + fe i & 5677.68 & 4.10 & @xmath302.640 & 7.52 & 7.02 & @xmath300.50 + fe i & 5678.60 & 2.42 & @xmath304.770 & 7.52 & 7.02 & @xmath300.50 + fe i & 5679.02 & 4.65 & @xmath300.900 & 7.52 & 7.02 & @xmath300.50 + fe i & 5680.24 & 4.19 & @xmath302.330 & 7.52 & 7.02 & @xmath300.50 + fe i & 5686.53 & 4.55 & @xmath300.626 & 7.52 & 7.02 & @xmath300.50 + fe i & 5691.50 & 4.30 & @xmath301.540 & 7.52 & 7.02 & @xmath300.50 + fe i & 5698.02 & 3.64 & @xmath302.790 & 7.52 & 7.02 & @xmath300.50 + fe i & 5699.41 & 4.96 & @xmath302.044 & 7.52 & 7.02 & @xmath300.50 + fe i & 5701.54 & 2.56 & @xmath302.046 & 7.52 & 7.02 & @xmath300.50 + fe i & 5704.73 & 5.03 & @xmath301.319 & 7.52 & 7.02 & @xmath300.50 + fe i & 5705.46 & 4.30 & @xmath301.565 & 7.52 & 7.02 & @xmath300.50 + fe i & 5707.70 & 4.10 & @xmath303.148 & 7.52 & 7.02 & @xmath300.50 + fe i & 5714.55 & 5.09 & @xmath301.715 & 7.52 & 7.02 & @xmath300.50 + fe i & 5715.47 & 4.15 & @xmath302.990 & 7.52 & 7.02 & @xmath300.50 + fe i & 5717.83 & 4.28 & @xmath301.090 & 7.52 & 7.02 & @xmath300.50 + fe i & 5720.89 & 4.55 & @xmath301.750 & 7.52 & 7.02 & @xmath300.50 + fe i & 5723.67 & 4.47 & @xmath302.250 & 7.52 & 7.02 & @xmath300.50 + fe i & 5724.45 & 4.28 & @xmath302.610 & 7.52 & 7.02 & @xmath300.50 + fe i & 5731.76 & 4.26 & @xmath301.210 & 7.52 & 7.02 & @xmath300.50 + fe i & 5732.30 & 4.99 & @xmath301.440 & 7.52 & 7.02 & @xmath300.50 + fe i & 5734.56 & 4.96 & @xmath301.784 & 7.52 & 7.02 & @xmath300.50 + fe i & 5738.23 & 4.22 & @xmath302.240 & 7.52 & 7.02 & @xmath300.50 + fe i & 5741.85 & 4.26 & @xmath301.744 & 7.52 & 7.02 & @xmath300.50 + fe i & 5750.03 & 5.01 & @xmath302.323 & 7.52 & 7.02 & @xmath300.50 + fe i & 5752.03 & 4.55 & @xmath301.077 & 7.52 & 7.02 & @xmath300.50 + fe i & 5759.26 & 4.65 & @xmath302.040 & 7.52 & 7.02 & @xmath300.50 + fe i & 5759.54 & 4.30 & @xmath302.179 & 7.52 & 7.02 & @xmath300.50 + fe i & 5760.34 & 3.64 & @xmath302.590 & 7.52 & 7.02 & @xmath300.50 + fe i & 5762.99 & 4.21 & @xmath300.460 & 7.52 & 7.02 & @xmath300.50 + fe i & 5767.97 & 4.29 & @xmath303.236 & 7.52 & 7.02 & @xmath300.50 + fe i & 5773.45 & 3.57 & @xmath303.704 & 7.52 & 7.02 & @xmath300.50 + fe i & 5775.08 & 4.22 & @xmath301.238 & 7.52 & 7.02 & @xmath300.50 + fe i & 5778.45 & 2.59 & @xmath303.590 & 7.52 & 7.02 & @xmath300.50 + fe i & 5784.66 & 3.40 & @xmath302.672 & 7.52 & 7.02 & @xmath300.50 + fe i & 5793.91 & 4.22 & @xmath301.750 & 7.52 & 7.02 & @xmath300.50 + fe i & 5809.22 & 3.88 & @xmath301.630 & 7.52 & 7.02 & @xmath300.50 + fe i & 5811.91 & 4.14 & @xmath302.460 & 7.52 & 7.02 & @xmath300.50 + fe i & 5814.81 & 4.28 & @xmath301.910 & 7.52 & 7.02 & @xmath300.50 + fe i & 5816.37 & 4.55 & @xmath300.681 & 7.52 & 7.02 & @xmath300.50 + fe i & 5821.89 & 4.99 & @xmath301.676 & 7.52 & 7.02 & @xmath300.50 + fe i & 5827.88 & 3.28 & @xmath303.260 & 7.52 & 7.02 & @xmath300.50 + fe i & 6120.25 & 0.92 & @xmath306.020 & 7.52 & 7.02 & @xmath300.50 + fe i & 6127.91 & 4.14 & @xmath301.499 & 7.52 & 7.02 & @xmath300.50 + fe i & 6136.61 & 2.45 & @xmath301.480 & 7.52 & 7.02 & @xmath300.50 + fe i & 6136.99 & 2.20 & @xmath302.900 & 7.52 & 7.02 & @xmath300.50 + fe i & 6137.69 & 2.59 & @xmath301.343 & 7.52 & 7.02 & @xmath300.50 + fe i & 6145.41 & 3.37 & @xmath303.770 & 7.52 & 7.02 & @xmath300.50 + fe i & 6151.62 & 2.18 & @xmath303.349 & 7.52 & 7.02 & @xmath300.50 + fe i & 6157.73 & 4.08 & @xmath301.110 & 7.52 & 7.02 & @xmath300.50 + fe i & 6159.37 & 4.61 & @xmath301.850 & 7.52 & 7.02 & @xmath300.50 + fe i & 6165.36 & 4.14 & @xmath301.614 & 7.52 & 7.02 & @xmath300.50 + fe i & 6171.01 & 4.73 & @xmath302.244 & 7.52 & 7.02 & @xmath300.50 + fe i & 6173.33 & 2.22 & @xmath302.870 & 7.52 & 7.02 & @xmath300.50 + fe i & 6180.20 & 2.73 & @xmath302.666 & 7.52 & 7.02 & @xmath300.50 + fe i & 6187.40 & 2.83 & @xmath304.168 & 7.52 & 7.02 & @xmath300.50 + fe i & 6187.99 & 3.94 & @xmath301.740 & 7.52 & 7.02 & @xmath300.50 + fe i & 6191.56 & 2.43 & @xmath301.367 & 7.52 & 7.02 & @xmath300.50 + fe i & 6200.31 & 2.61 & @xmath302.407 & 7.52 & 7.02 & @xmath300.50 + fe i & 6213.43 & 2.22 & @xmath302.542 & 7.52 & 7.02 & @xmath300.50 + fe i & 6219.28 & 2.20 & @xmath302.353 & 7.52 & 7.02 & @xmath300.50 + fe i & 6226.73 & 3.88 & @xmath302.210 & 7.52 & 7.02 & @xmath300.50 + fe i & 6229.23 & 2.85 & @xmath302.955 & 7.52 & 7.02 & @xmath300.50 + fe i & 6232.64 & 3.65 & @xmath301.323 & 7.52 & 7.02 & @xmath300.50 + fe i & 6240.65 & 2.22 & @xmath303.333 & 7.52 & 7.02 & @xmath300.50 + fe i & 6246.32 & 3.60 & @xmath300.953 & 7.52 & 7.02 & @xmath300.50 + fe i & 6252.56 & 2.40 & @xmath301.697 & 7.52 & 7.02 & @xmath300.50 + fe i & 6270.22 & 2.86 & @xmath302.704 & 7.52 & 7.02 & @xmath300.50 + fe i & 6322.69 & 2.59 & @xmath302.376 & 7.52 & 7.02 & @xmath300.50 + fe i & 6330.85 & 4.73 & @xmath301.290 & 7.52 & 7.02 & @xmath300.50 + fe i & 6335.33 & 2.20 & @xmath302.187 & 7.52 & 7.02 & @xmath300.50 + fe i & 6336.82 & 3.69 & @xmath300.966 & 7.52 & 7.02 & @xmath300.50 + fe i & 6380.74 & 4.19 & @xmath301.326 & 7.52 & 7.02 & @xmath300.50 + fe i & 6392.54 & 2.28 & @xmath304.090 & 7.52 & 7.02 & @xmath300.50 + fe i & 6393.60 & 2.43 & @xmath301.562 & 7.52 & 7.02 & @xmath300.50 + fe i & 6608.02 & 2.28 & @xmath304.070 & 7.52 & 7.02 & @xmath300.50 + fe i & 6609.11 & 2.56 & @xmath302.602 & 7.52 & 7.02 & @xmath300.50 + fe i & 6609.68 & 0.99 & @xmath305.700 & 7.52 & 7.02 & @xmath300.50 + fe i & 6648.08 & 1.01 & @xmath305.824 & 7.52 & 7.02 & @xmath300.50 + fe i & 6699.14 & 4.59 & @xmath302.081 & 7.52 & 7.02 & @xmath300.50 + fe i & 6705.10 & 4.61 & @xmath301.122 & 7.52 & 7.02 & @xmath300.50 + fe i & 6710.32 & 1.49 & @xmath304.890 & 7.52 & 7.02 & @xmath300.50 + fe i & 6713.74 & 4.79 & @xmath301.530 & 7.52 & 7.02 & @xmath300.50 + fe i & 6726.67 & 4.61 & @xmath301.183 & 7.52 & 7.02 & @xmath300.50 + fe i & 6733.15 & 4.64 & @xmath301.550 & 7.52 & 7.02 & @xmath300.50 + fe i & 6737.27 & 3.27 & @xmath304.339 & 7.52 & 7.02 & @xmath300.50 + fe i & 6750.15 & 2.42 & @xmath302.681 & 7.52 & 7.02 & @xmath300.50 + fe i & 6806.84 & 2.73 & @xmath303.180 & 7.52 & 7.02 & @xmath300.50 + fe i & 6810.26 & 4.61 & @xmath301.086 & 7.52 & 7.02 & @xmath300.50 + fe i & 6820.37 & 4.64 & @xmath301.130 & 7.52 & 7.02 & @xmath300.50 + fe i & 6842.69 & 4.64 & @xmath301.270 & 7.52 & 7.02 & @xmath300.50 + fe i & 6843.65 & 4.55 & @xmath300.960 & 7.52 & 7.02 & @xmath300.50 + fe i & 6857.25 & 4.08 & @xmath302.230 & 7.52 & 7.02 & @xmath300.50 + fe
i & 6864.31 & 4.56 & @xmath302.410 & 7.52 & 7.02 & @xmath300.50 + fe ii & 6149.26 & 3.89 & @xmath302.681 & 7.52 & 7.02 & @xmath300.50 + fe ii & 6247.56 & 3.89 & @xmath302.405 & 7.52 & 7.02 & @xmath300.50 + fe ii & 6369.46 & 2.89 & @xmath304.141 & 7.52 & 7.02 & @xmath300.50 + co i & 5647.23 & 2.28 & hfs & 4.90 & 4.52 & @xmath00.12 + co i & 6117.00 & 1.78 & hfs & 4.90 & 4.52 & @xmath00.12 + ni i & 5593.73 & 3.90 & @xmath300.960 & 6.25 & 5.81 & @xmath00.06 + ni i & 5614.77 & 4.15 & @xmath300.698 & 6.25 & 5.81 & @xmath00.06 + ni i & 5625.31 & 4.09 & @xmath300.750 & 6.25 & 5.81 & @xmath00.06 + ni i & 5628.34 & 4.09 & @xmath301.301 & 6.25 & 5.81 & @xmath00.06 + ni i & 5638.74 & 3.90 & @xmath301.670 & 6.25 & 5.81 & @xmath00.06 + ni i & 5641.88 & 4.11 & @xmath301.080 & 6.25 & 5.81 & @xmath00.06 + ni i & 5643.07 & 4.17 & @xmath301.260 & 6.25 & 5.81 & @xmath00.06 + ni i & 5682.20 & 4.11 & @xmath300.510 & 6.25 & 5.81 & @xmath00.06 + ni i & 5694.98 & 4.09 & @xmath300.760 & 6.25 & 5.81 & @xmath00.06 + ni i & 5748.35 & 1.68 & @xmath303.160 & 6.25 & 5.81 & @xmath00.06 + ni i & 5760.83 & 4.11 & @xmath300.790 & 6.25 & 5.81 & @xmath00.06 + ni i & 5796.08 & 1.95 & @xmath303.752 & 6.25 & 5.81 & @xmath00.06 + ni i & 5805.21 & 4.17 & @xmath300.720 & 6.25 & 5.81 & @xmath00.06 + ni i & 6128.96 & 1.68 & @xmath303.400 & 6.25 & 5.81 & @xmath00.06 + ni i & 6130.13 & 4.27 & @xmath301.040 & 6.25 & 5.81 & @xmath00.06 + ni i & 6175.36 & 4.09 & @xmath300.619 & 6.25 & 5.81 & @xmath00.06 + ni i & 6176.81 & 4.09 & @xmath300.270 & 6.25 & 5.81 & @xmath00.06 + ni i & 6177.24 & 1.83 & @xmath303.550 & 6.25 & 5.81 & @xmath00.06 + ni i & 6186.71 & 4.11 & @xmath300.890 & 6.25 & 5.81 & @xmath00.06 + ni i & 6191.17 & 1.68 & @xmath302.233 & 6.25 & 5.81 & @xmath00.06 + ni i & 6223.98 & 4.11 & @xmath300.960 & 6.25 & 5.81 & @xmath00.06 + ni i & 6322.16 & 4.15 & @xmath301.190 & 6.25 & 5.81 & @xmath00.06 + ni i & 6635.12 & 4.42 & @xmath300.750 & 6.25 & 5.81 & @xmath00.06 + ni i & 6767.77 & 1.83 & @xmath302.100 & 6.25 & 5.81 & @xmath00.06 + ni i & 6772.31 & 3.66 & @xmath301.010 & 6.25 & 5.81 & @xmath00.06
+ ni i & 6813.60 & 5.34 & @xmath300.354 & 6.25 & 5.81 & @xmath00.06 + cu i & 5782.11 & 1.64 & hfs & 4.04 & 3.71 & @xmath00.17 + + 119799c4 & @xmath00.63 & @xmath00.08 & @xmath00.29 & @xmath00.26 & @xmath00.38 & @xmath00.17 & @xmath00.07 & @xmath300.39 & @xmath300.38 & @xmath00.25 & @xmath00.21 & @xmath00.26 + 129499c4 & & @xmath00.25 & @xmath00.22 & @xmath00.24 & @xmath00.04 & @xmath00.30 & @xmath00.26 & @xmath00.13 & & @xmath00.32 & @xmath00.06 & + 176772c5 & @xmath00.12 & @xmath00.19 & @xmath00.15 & @xmath00.08 & @xmath300.05 & @xmath00.16 & @xmath00.05 & @xmath00.07 & @xmath00.05 & @xmath00.07 & @xmath00.14 & + 181349c5 & @xmath300.06 & @xmath00.58 & @xmath00.15 & @xmath00.20 & @xmath00.11 & @xmath00.08 & @xmath00.00 & @xmath00.30 & @xmath00.21 & @xmath00.25 & @xmath00.10 & + 183783c5 & & @xmath00.32 & @xmath00.12 & @xmath00.30 & @xmath300.11 & @xmath00.15 & @xmath00.15 & @xmath00.10 & & @xmath00.32 & @xmath00.08 & + 184088c5 & @xmath00.74 & @xmath00.03 & @xmath00.30 & @xmath00.37 & @xmath00.14 & @xmath00.37 & @xmath00.09 & @xmath300.45 & @xmath300.48 & @xmath00.37 & @xmath00.15 & @xmath00.37 + 184618c5 & & @xmath00.00 & @xmath00.63 & @xmath00.28 & @xmath00.24 & @xmath00.31 & @xmath00.06 & @xmath300.58 & @xmath300.59 & @xmath00.14 & @xmath00.15 & @xmath00.28 + 185169c5 & & @xmath00.22 & @xmath00.37 & @xmath00.42 & @xmath00.12 & @xmath00.53 & @xmath00.21 & @xmath300.32 & & @xmath00.38 & @xmath00.07 & @xmath00.39 + 185357c5 & @xmath300.05 & @xmath00.24 & @xmath00.01 & @xmath00.06 & @xmath300.11 & @xmath300.04 & @xmath300.09 & @xmath00.29 & @xmath00.27 & @xmath00.04 & @xmath00.12 & + 185541c5 & @xmath300.05 & @xmath00.21 & @xmath300.01 & @xmath00.01 & @xmath00.06 & @xmath300.06 & @xmath00.08 & @xmath00.34 & @xmath00.33 & @xmath00.18 & @xmath00.06 & + 187067c5 & & @xmath300.35 & @xmath00.30 & @xmath300.10 & @xmath00.19 & @xmath00.53 & @xmath300.03 & @xmath300.95 & @xmath300.94 & @xmath00.07 & @xmath00.07 & @xmath300.43 + 193190c5 & & @xmath00.37 & @xmath00.03 & @xmath00.08 & @xmath00.04 & @xmath00.25 & @xmath00.16 & @xmath00.25 & @xmath00.22 & @xmath00.22 & @xmath00.06 & + 197366c5 & @xmath00.01 & @xmath00.48 & @xmath300.08 & @xmath00.12 & @xmath300.05 & @xmath00.13 & @xmath00.12 & @xmath00.33 & @xmath00.31 & @xmath00.26 & @xmath00.19 & + 215681c6 & @xmath00.18 & @xmath00.33 & @xmath00.09 & @xmath00.06 & @xmath300.02 & @xmath00.21 & @xmath00.07 & @xmath00.31 & & @xmath00.20 & @xmath00.05 & @xmath00.51 + 216922c6 & @xmath00.43 & @xmath00.28 & @xmath00.29 & @xmath00.34 & @xmath00.24 & @xmath00.42 & @xmath00.25 & @xmath300.19 & @xmath300.20 & @xmath00.24 & @xmath00.08 & + 218198c6 & @xmath00.50 & @xmath00.32 & @xmath00.31 & @xmath00.27 & @xmath00.30 & @xmath00.32 & @xmath00.16 & @xmath300.31 & @xmath300.33 & @xmath00.23 & @xmath00.16 & + 219909c6 & @xmath00.52 & @xmath300.20 & @xmath00.38 & @xmath00.28 & @xmath00.25 & @xmath00.32 & @xmath300.02 & @xmath300.73 & @xmath300.73 & @xmath00.20 & @xmath00.04 & + 221537c6 & @xmath300.35 & @xmath00.31 & @xmath00.01 & @xmath00.09 & @xmath300.20 & @xmath00.14 & @xmath00.12 & @xmath00.29 & & @xmath00.08 & @xmath00.06 & + 223113c6 & @xmath00.43 & @xmath00.01 & @xmath00.29 & @xmath00.12 & @xmath00.14 & @xmath00.25 & @xmath00.01 & @xmath300.14 & @xmath300.15 & @xmath00.15 & @xmath00.08 & + 223343c6 & @xmath00.11 & @xmath00.00 & @xmath300.10 & @xmath00.10 & @xmath00.12 & @xmath00.16 & @xmath300.05 & @xmath00.15 & @xmath00.14 & @xmath00.15 & @xmath00.10 & + 223621c6 & & @xmath300.04 & @xmath00.02 & @xmath00.09 & @xmath300.12 & @xmath00.25 & @xmath00.07 & @xmath00.33 & @xmath00.28 & @xmath00.34 & @xmath00.07 & + 223722c6 & @xmath00.40 & @xmath300.05 & @xmath00.44 & @xmath00.28 & @xmath00.33 & @xmath00.43 & @xmath300.05 & @xmath300.76 & @xmath300.74 & @xmath300.01 & @xmath00.02 & + 224206c6 & & @xmath00.10 & @xmath300.10 & @xmath300.05 & @xmath300.03 & @xmath300.24 & @xmath00.09 & @xmath00.46 & @xmath00.45 & @xmath00.21 & @xmath00.11 & + 224866c6 & @xmath00.29 & @xmath00.17 & @xmath00.22 & @xmath00.31 & @xmath00.16 & @xmath00.25 & @xmath300.03 & @xmath300.22 & @xmath300.21 & @xmath00.17 & @xmath00.11 & + 224951c6 & @xmath00.32 & @xmath00.22 & @xmath00.07 & @xmath00.05 & @xmath00.19 & @xmath00.14 & @xmath00.04 & @xmath00.17 & & @xmath00.14 & @xmath00.05 & + 225531c6 & @xmath00.66 & @xmath00.20 & @xmath00.46 & @xmath00.31 & @xmath00.35 & @xmath00.40 & @xmath00.02 & @xmath300.72 & @xmath300.73 & @xmath00.29 & @xmath00.10 & @xmath00.28 + 226450c6 & @xmath300.16 & @xmath00.06 & @xmath00.03 & @xmath00.03 & @xmath00.09 & @xmath00.07 & @xmath300.13 & @xmath00.27 & & @xmath00.08 & @xmath00.05 & + 226850c6 & & @xmath300.25 & @xmath00.35 & @xmath00.18 & @xmath00.33 & @xmath00.35 & @xmath00.04 & @xmath300.85 & @xmath300.84 & @xmath00.15 & @xmath00.00 & @xmath300.18 + 227867c6 & @xmath00.51 & @xmath300.08 & @xmath00.12 & @xmath00.06 & @xmath00.11 & @xmath00.08 & @xmath300.05 & @xmath00.11 & @xmath00.07 & @xmath00.17 & @xmath00.14 & @xmath00.38 + 228466c6 & @xmath00.55 & @xmath300.22 & @xmath00.36 & @xmath00.34 & @xmath00.29 & @xmath00.32 & @xmath00.31 & @xmath300.96 & @xmath300.99 & @xmath00.18 & @xmath00.06 & @xmath00.03 + 229507c6 & & @xmath00.19 & @xmath00.15 & @xmath00.20 & @xmath00.06 & @xmath00.26 & @xmath300.14 & @xmath300.30 & @xmath300.28 & @xmath00.00 & @xmath00.11 & @xmath00.15 + 230424c6 & @xmath00.57 & @xmath300.19 & @xmath00.38 & @xmath00.18 & @xmath00.28 & @xmath00.24 & @xmath300.02 & @xmath300.83 & @xmath300.85 & @xmath00.19 & @xmath00.13 & @xmath00.20 + 230483c6 & @xmath00.62 & @xmath00.13 & @xmath00.35 & @xmath00.42 & @xmath00.32 & @xmath00.31 & @xmath00.22 & @xmath300.63 & @xmath300.66 & @xmath00.33 & @xmath00.13 & + 231379c6 & @xmath00.54 & @xmath00.15 & @xmath00.39 & @xmath00.28 & @xmath00.37 & @xmath00.32 & @xmath00.01 & @xmath300.30 & @xmath300.29 & @xmath00.32 & @xmath00.16 & @xmath00.45 + 231618c6 & @xmath300.35 & @xmath300.07 & @xmath00.09 & @xmath00.06 & @xmath00.12 & @xmath00.05 & @xmath300.03 & @xmath00.04 & @xmath00.04 & @xmath00.04 & @xmath00.11 & @xmath00.28 + 232493c6 & @xmath00.68 & @xmath300.26 & @xmath00.29 & @xmath00.14 & @xmath00.04 & @xmath00.45 & @xmath300.31 & @xmath301.14 & @xmath301.13 & @xmath300.03 & @xmath300.16 & @xmath300.55 + 233121c6 & @xmath00.63 & @xmath00.01 & @xmath00.31 & @xmath00.26 & @xmath00.33 & @xmath00.29 & @xmath00.04 & @xmath300.49 & @xmath300.50 & @xmath00.16 & @xmath00.07 & + 233560c6 & @xmath00.44 & @xmath300.05 & @xmath00.30 & @xmath00.14 & @xmath00.34 & @xmath00.25 & @xmath300.13 & @xmath300.20 & @xmath300.22 & @xmath00.19 & @xmath00.16 & @xmath00.40 + 233708c6 & @xmath300.13 & @xmath00.15 & @xmath00.03 & @xmath300.01 & @xmath00.13 & @xmath00.14 & @xmath300.11 & @xmath00.12 & @xmath00.16 & @xmath00.09 & @xmath00.17 & @xmath00.15 + 240059c6 & @xmath00.64 & @xmath300.11 & @xmath00.45 & @xmath00.27 & @xmath00.29 & @xmath00.45 & @xmath300.03 & @xmath300.70 & @xmath300.68 & @xmath00.11 & @xmath00.02 & + 240083c6 & @xmath00.38 & @xmath300.08 & @xmath00.35 & @xmath00.27 & @xmath00.24 & @xmath00.34 & @xmath300.02 & @xmath300.59 & @xmath300.56 & @xmath00.04 & @xmath00.00 & + 259050c7 & @xmath00.44 & @xmath300.08 & @xmath00.40 & @xmath00.22 & @xmath00.35 & @xmath00.13 & @xmath300.04 & @xmath300.65 & @xmath300.65 & @xmath00.04 & @xmath00.12 & + 259377c7 & & @xmath00.09 & @xmath300.03 & @xmath00.03 & @xmath00.00 & @xmath00.08 & @xmath00.03 & @xmath00.29 & & @xmath00.26 & @xmath00.08 & @xmath00.78 + 260308c7 & @xmath00.24 & @xmath00.69 & @xmath00.04 & @xmath00.22 & @xmath00.02 & @xmath00.20 & @xmath300.02 & @xmath00.15 & @xmath00.16 & @xmath00.19 & @xmath00.16 & + 262018c7 & @xmath00.34 & @xmath300.02 & @xmath00.12 & @xmath00.00 & @xmath00.00 & @xmath00.01 & @xmath300.02 & @xmath00.08 & @xmath00.07 & @xmath00.15 & @xmath300.04 & @xmath00.46 + 266442c7 & @xmath00.56 & @xmath00.22 & @xmath00.17 & @xmath00.34 & @xmath00.29 & @xmath00.35 & @xmath00.02 & @xmath300.42 & @xmath300.42 & @xmath00.24 & @xmath00.15 & @xmath00.59 + 270316c7 & & @xmath00.23 & @xmath00.19 & @xmath00.32 & @xmath00.34 & @xmath00.42 & @xmath00.01 & @xmath300.37 & @xmath300.37 & @xmath00.14 & @xmath00.19 & @xmath00.04 + 275181c7 & @xmath00.53 & @xmath00.12 & @xmath00.29 & @xmath00.39 & @xmath00.34 & @xmath00.53 & @xmath00.17 & @xmath300.59 & @xmath300.59 & @xmath00.21 & @xmath00.09 & @xmath00.36 + 277490c7 & @xmath00.45 & @xmath300.07 & @xmath00.31 & @xmath00.30 & @xmath00.33 & @xmath00.37 & @xmath300.05 & @xmath300.56 & @xmath300.55 & @xmath00.17 & @xmath00.12 & @xmath00.00 + 278419c7 & @xmath300.36 & @xmath00.18 & @xmath00.05 & @xmath00.14 & @xmath00.00 & @xmath00.28 & @xmath300.01 & @xmath00.20 & & @xmath00.22 & @xmath00.08 & + 282804c7 & @xmath300.21 & @xmath00.26 & @xmath300.03 & @xmath00.22 & @xmath00.07 & @xmath00.21 & @xmath00.04 & @xmath00.10 & & @xmath00.11 & @xmath00.11 & @xmath00.25 + 286252c7 & & @xmath00.10 & @xmath300.15 & @xmath00.01 & @xmath00.02 & @xmath00.07 & @xmath300.16 & @xmath00.42 & @xmath00.44 & @xmath00.18 & @xmath00.21 & @xmath00.77 + 45512c2 & @xmath00.64 & @xmath300.11 & @xmath00.21 & @xmath00.10 & @xmath00.36 & @xmath00.12 & @xmath300.21 & @xmath300.30 & @xmath300.23 & @xmath00.11 & @xmath00.20 & @xmath00.52 + 47188c2 & @xmath00.71 & @xmath300.08 & @xmath00.38 & @xmath00.17 & @xmath00.37 & @xmath00.33 & @xmath300.06 & @xmath300.52 & & @xmath00.11 & @xmath00.10 & + 77186c3 & @xmath300.07 & @xmath00.10 & @xmath300.03 & @xmath300.11 & @xmath00.07 & @xmath300.06 & @xmath00.07 & @xmath00.41 & @xmath00.41 & @xmath00.18 & @xmath00.13 & @xmath00.55 + 77707c3 & @xmath300.30 & @xmath00.00 & @xmath300.04 & @xmath300.03 & @xmath300.20 & @xmath300.02 & @xmath00.06 & @xmath00.44 & @xmath00.43 & @xmath00.23 & @xmath00.10 & @xmath00.48 + 80582c3 & & @xmath00.06 & @xmath00.21 & @xmath00.03 & @xmath00.08 & @xmath00.11 & @xmath00.03 & @xmath00.07 & @xmath00.06 & @xmath00.28 & @xmath00.12 & @xmath00.79 + 81644c3 & @xmath300.08 & @xmath00.13 & @xmath00.03 & @xmath00.15 & @xmath00.09 & @xmath00.11 & @xmath00.10 & @xmath00.22 & @xmath00.20 & @xmath00.15 & @xmath00.09 & @xmath00.33 + 82227c3 & @xmath300.06 & @xmath00.05 & @xmath00.03 & @xmath00.09 & @xmath00.11 & @xmath00.03 & @xmath300.17 & @xmath00.25 & & @xmath00.15 & @xmath00.21 & @xmath00.57 + 83531c3 & @xmath300.06 & @xmath00.04 & @xmath00.06 & @xmath300.20 & @xmath00.02 & @xmath300.04 & @xmath300.07 & @xmath00.45 & @xmath00.44 & @xmath00.25 & @xmath00.16 & @xmath00.77 + 84255c3 & @xmath00.36 & @xmath300.24 & @xmath00.34 & @xmath00.16 & @xmath00.20 & @xmath00.55 & @xmath300.03 & @xmath300.70 & @xmath300.66 & @xmath300.05 & @xmath300.01 & @xmath300.03 + 86757c3 & & @xmath00.02 & @xmath00.04 & @xmath00.02 & @xmath300.05 & @xmath00.03 & @xmath00.01 & @xmath300.02 & @xmath300.02 & @xmath00.03 & @xmath00.09 & @xmath00.39 + 88522c3 & & @xmath00.09 & @xmath300.02 & @xmath300.05 & @xmath00.06 & @xmath00.03 & @xmath00.02 & @xmath00.37 & @xmath00.40 & @xmath00.14 & @xmath00.09 & @xmath00.50 + + 225847c6 & & @xmath300.03 & @xmath00.10 & @xmath00.00 & @xmath00.26 & @xmath00.02 & @xmath300.18 & @xmath300.11 & @xmath300.10 & @xmath300.06 & @xmath00.10 & + 227379c6 & @xmath00.24 & @xmath00.30 & @xmath00.23 & @xmath00.04 & @xmath00.36 & @xmath00.15 & @xmath300.08 & @xmath300.05 & @xmath300.06 & @xmath00.12 & @xmath00.07 & + 228407c6 & & @xmath00.45 & @xmath00.19 & @xmath00.28 & @xmath00.22 & @xmath00.28 & @xmath00.08 & @xmath300.21 & @xmath300.23 & @xmath00.11 & @xmath00.07 & @xmath00.38 + 230208c6 & & @xmath300.26 & @xmath00.07 & @xmath300.02 & @xmath00.04 & @xmath00.04 & @xmath300.14 & @xmath00.01 & @xmath300.01 & @xmath00.00 & @xmath300.01 & @xmath00.31 + 239284c6 & & @xmath00.17 & @xmath00.23 & @xmath00.27 & @xmath00.16 & @xmath00.35 & @xmath00.12 & @xmath300.14 & & @xmath00.16 & @xmath00.14 & + 265795c7 & & @xmath00.17 & @xmath00.02 & @xmath00.21 & @xmath00.04 & @xmath00.40 & @xmath00.11 & @xmath300.07 & & @xmath00.21 & @xmath00.04 & @xmath00.84 + 268360c7 & & @xmath00.00 & @xmath00.11 & @xmath00.23 & @xmath00.10 & @xmath00.22 & @xmath00.04 & @xmath300.04 & @xmath300.12 & @xmath00.31 & @xmath00.16 & @xmath00.51 + 268493c7 & & @xmath00.46 & @xmath00.24 & @xmath00.35 & @xmath00.27 & @xmath00.20 & @xmath300.05 & @xmath300.19 & @xmath300.20 & @xmath00.20 & @xmath00.05 & @xmath00.49 + 271021c7 & & @xmath00.12 & @xmath00.09 & @xmath00.01 & @xmath00.11 & @xmath00.29 & @xmath00.19 & @xmath300.07 & @xmath300.05 & @xmath00.15 & @xmath00.09 & @xmath00.26 + 271400c7 & & @xmath00.22 & @xmath00.24 & @xmath00.30 & @xmath00.16 & @xmath00.32 & @xmath00.14 & @xmath300.18 & @xmath300.20 & @xmath00.25 & @xmath00.06 & @xmath00.45 + 77182c3 & & @xmath00.18 & @xmath00.21 & @xmath00.11 & @xmath00.07 & @xmath00.19 & @xmath300.01 & @xmath300.09 & @xmath300.09 & @xmath00.13 & @xmath00.14 & + 85597c3 & & @xmath00.13 & @xmath00.20 & @xmath00.22 & @xmath00.23 & @xmath00.18 & @xmath300.07 & @xmath300.15 & & @xmath00.10 & @xmath00.10 & @xmath00.50 + + 1156c2 & @xmath00.31 & @xmath00.06 & @xmath00.35 & @xmath00.24 & @xmath00.35 & @xmath00.27 & @xmath300.06 & @xmath300.45 & @xmath300.46 & @xmath00.16 & @xmath00.09 & @xmath00.22 + 1407c3 & @xmath00.55 & @xmath300.06 & @xmath00.26 & @xmath00.10 & @xmath00.23 & @xmath00.38 & @xmath00.09 & @xmath300.61 & @xmath300.62 & @xmath00.05 & @xmath300.07 & @xmath300.06 + 1491c7 & @xmath00.56 & @xmath00.19 & @xmath00.37 & @xmath00.32 & @xmath00.26 & @xmath00.38 & @xmath00.13 & @xmath300.32 & & @xmath00.32 & @xmath00.09 & @xmath00.19 + 1554c7 & @xmath00.61 & @xmath00.05 & @xmath00.42 & @xmath00.37 & @xmath00.24 & @xmath00.44 & @xmath00.25 & @xmath300.67 & & @xmath00.17 & @xmath00.01 & @xmath300.01 + 166c3 & @xmath00.56 & @xmath300.18 & @xmath00.42 & @xmath00.48 & @xmath00.20 & @xmath00.46 & @xmath00.04 & @xmath300.82 & @xmath300.83 & @xmath00.17 & @xmath300.09 & @xmath300.19 + 1754c3 & & @xmath00.03 & @xmath00.29 & @xmath00.12 & @xmath00.26 & @xmath00.34 & @xmath00.00 & @xmath300.22 & @xmath300.28 & @xmath00.22 & @xmath00.06 & @xmath00.11 + 1814c1 & @xmath00.57 & @xmath300.05 & @xmath00.28 & @xmath00.20 & @xmath00.25 & @xmath00.27 & @xmath300.04 & @xmath300.30 & @xmath300.29 & @xmath00.25 & @xmath00.13 & @xmath00.37 + 1876c2 & & @xmath300.18 & @xmath00.40 & @xmath00.17 & @xmath00.43 & @xmath00.44 & @xmath300.14 & @xmath301.02 & @xmath301.04 & @xmath00.02 & @xmath300.07 & @xmath300.41 + 1917c1 & & @xmath00.20 & @xmath300.07 & @xmath00.02 & @xmath300.07 & @xmath00.09 & @xmath00.03 & @xmath00.20 & @xmath00.16 & @xmath00.13 & @xmath00.14 & @xmath00.37 + 1918c1 & @xmath00.49 & @xmath300.02 & @xmath00.35 & @xmath00.23 & @xmath00.28 & @xmath00.29 & @xmath300.04 & @xmath300.43 & @xmath300.44 & @xmath00.13 & @xmath00.10 & @xmath300.01 + 201583c3 & & @xmath300.09 & @xmath00.38 & @xmath00.11 & @xmath00.15 & @xmath00.50 & & @xmath301.11 & @xmath301.10 & & @xmath00.10 & @xmath300.77 + 2110c7 & @xmath00.33 & @xmath300.05 & @xmath00.12 & @xmath00.15 & @xmath00.17 & @xmath00.18 & @xmath300.04 & @xmath00.01 & @xmath00.02 & @xmath00.15 & @xmath00.09 & @xmath00.52 + 2178c7 & @xmath00.20 & @xmath300.06 & @xmath00.17 & @xmath00.12 & @xmath00.17 & @xmath300.01 & @xmath300.19 & @xmath00.01 & @xmath300.01 & @xmath00.06 & @xmath00.09 & @xmath00.15 + 2200c3 & & @xmath300.01 & @xmath00.26 & @xmath00.15 & @xmath00.20 & @xmath00.16 & @xmath00.08 & @xmath300.10 & @xmath300.10 & @xmath00.15 & @xmath00.11 & @xmath00.32 + 2220c7 & @xmath00.23 & @xmath00.24 & @xmath00.36 & @xmath00.39 & @xmath00.03 & @xmath00.53 & @xmath00.33 & @xmath300.17 & & @xmath00.24 & @xmath00.02 & @xmath00.32 + 222c3 & & @xmath300.01 & @xmath00.36 & @xmath00.32 & @xmath00.28 & @xmath00.23 & @xmath00.10 & @xmath300.42 & @xmath300.43 & @xmath00.22 & @xmath00.09 & @xmath00.19 + 2335c2 & @xmath00.47 & @xmath300.20 & @xmath00.34 & @xmath00.26 & @xmath00.29 & @xmath00.39 & @xmath300.03 & @xmath300.91 & @xmath300.89 & @xmath00.04 & @xmath00.03 & @xmath300.24 + 2407c2 & @xmath00.38 & @xmath300.13 & @xmath00.24 & @xmath00.23 & @xmath00.37 & @xmath00.40 & @xmath300.15 & @xmath300.64 & @xmath300.63 & @xmath00.01 & @xmath00.05 & @xmath300.14 + 2422c7 & @xmath300.18 & @xmath00.05 & @xmath00.05 & @xmath300.07 & @xmath00.20 & @xmath300.11 & @xmath300.16 & @xmath00.17 & & @xmath00.00 & @xmath00.08 & @xmath00.15 + 2470c3 & @xmath00.14 & @xmath00.06 & @xmath00.27 & @xmath00.16 & @xmath00.09 & @xmath00.17 & @xmath00.04 & @xmath00.00 & @xmath300.01 & @xmath00.22 & @xmath00.10 & @xmath00.37 + 2502c3 & @xmath00.52 & @xmath00.05 & @xmath00.30 & @xmath00.21 & @xmath00.27 & @xmath00.38 & @xmath00.14 & @xmath300.63 & @xmath300.61 & @xmath00.10 & @xmath00.02 & @xmath300.10 + 2532c6 & @xmath00.59 & @xmath00.02 & @xmath00.36 & @xmath00.20 & @xmath00.23 & @xmath00.19 & @xmath300.05 & @xmath300.18 & @xmath300.16 & @xmath00.25 & @xmath00.08 & @xmath00.55 + 2580c6 & @xmath00.04 & @xmath00.23 & @xmath00.13 & @xmath00.05 & @xmath00.25 & @xmath300.05 & @xmath00.13 & @xmath00.30 & @xmath00.30 & @xmath00.24 & @xmath00.23 & @xmath00.47 + 2769c3 & @xmath00.03 & @xmath00.07 & @xmath00.17 & @xmath00.14 & @xmath00.06 & @xmath00.25 & @xmath00.07 & @xmath00.06 & @xmath00.06 & @xmath00.06 & @xmath00.09 & @xmath00.21 + 2772c7 & @xmath00.69 & @xmath00.11 & @xmath00.25 & @xmath00.30 & @xmath00.29 & @xmath00.23 & @xmath300.03 & @xmath300.25 & @xmath300.24 & @xmath00.25 & @xmath00.10 & @xmath00.37 + 2812c8 & @xmath00.45 & @xmath300.11 & @xmath00.30 & @xmath00.29 & @xmath00.33 & @xmath00.30 & @xmath300.02 & @xmath300.64 & @xmath300.65 & @xmath00.08 & @xmath00.08 & @xmath00.16 + 2947c3 & @xmath00.65 & @xmath300.08 & @xmath00.24 & @xmath00.26 & @xmath00.24 & @xmath00.25 & @xmath00.02 & @xmath300.49 & @xmath300.50 & @xmath00.24 & @xmath00.08 & @xmath00.24 + 2948c7 & @xmath300.38 & @xmath00.28 & @xmath00.03 & @xmath00.03 & @xmath00.04 & @xmath00.13 & @xmath300.01 & @xmath00.12 & & @xmath300.04 & @xmath00.15 & @xmath00.05 + 3018c3 & @xmath00.52 & @xmath300.10 & @xmath00.28 & @xmath00.22 & @xmath00.26 & @xmath00.30 & @xmath300.06 & @xmath300.53 & @xmath300.55 & @xmath00.15 & @xmath00.05 & @xmath00.15 + 3035c7 & @xmath00.65 & @xmath300.24 & @xmath00.26 & @xmath00.09 & @xmath00.27 & @xmath00.34 & @xmath300.13 & @xmath300.44 & @xmath300.43 & @xmath00.06 & @xmath00.02 & @xmath300.07 + 3091c8 & @xmath00.53 & @xmath00.00 & @xmath00.38 & @xmath00.26 & @xmath00.25 & @xmath00.29 & @xmath00.00 & @xmath300.49 & @xmath300.50 & @xmath00.20 & @xmath00.13 & @xmath00.11 + 3101c7 & @xmath00.50 & @xmath300.07 & @xmath00.37 & @xmath00.21 & @xmath00.32 & @xmath00.22 & @xmath300.05 & @xmath300.39 & @xmath300.40 & @xmath00.22 & @xmath00.09 & @xmath00.07 + 3142c3 & @xmath00.51 & @xmath300.09 & @xmath00.27 & @xmath00.19 & @xmath00.25 & @xmath00.16 & @xmath00.00 & @xmath300.22 & @xmath300.21 & @xmath00.13 & @xmath00.09 & @xmath00.21 + 3161c3 & & @xmath00.16 & @xmath00.29 & @xmath00.31 & @xmath00.08 & @xmath00.29 & @xmath00.15 & @xmath300.14 & @xmath300.16 & @xmath00.34 & @xmath00.07 & @xmath00.41 + 3191c7 & @xmath00.55 & @xmath300.06 & @xmath00.37 & @xmath00.21 & @xmath00.24 & @xmath00.29 & @xmath300.01 & @xmath300.36 & @xmath300.37 & @xmath00.13 & @xmath00.09 & @xmath00.09 + 3201c6 & @xmath00.01 & @xmath00.04 & @xmath00.12 & @xmath00.19 & @xmath00.03 & @xmath00.25 & @xmath00.14 & @xmath00.03 & @xmath00.04 & @xmath00.12 & @xmath00.04 & @xmath00.22 + 3238c6 & & @xmath300.01 & @xmath00.19 & @xmath00.18 & @xmath00.20 & @xmath00.23 & @xmath300.03 & @xmath300.21 & @xmath300.22 & @xmath00.24 & @xmath00.11 & @xmath00.32 + 3267c3 & @xmath300.13 & @xmath300.15 & @xmath00.09 & @xmath300.02 & @xmath00.14 & @xmath00.05 & @xmath300.19 & @xmath00.04 & @xmath00.03 & @xmath00.01 & @xmath00.09 & @xmath00.18 + 3515c3 & & @xmath00.27 & @xmath00.19 & @xmath00.15 & @xmath00.13 & @xmath00.26 & @xmath00.08 & @xmath00.01 & @xmath00.01 & @xmath00.11 & @xmath00.05 & @xmath00.30 + 3558c6 & & @xmath00.01 & @xmath00.33 & @xmath00.20 & @xmath00.19 & @xmath00.16 & @xmath300.15 & @xmath300.15 & @xmath300.14 & @xmath00.22 & @xmath00.11 & @xmath00.57 + 3690c7 & @xmath00.47 & @xmath300.32 & @xmath00.28 & @xmath00.38 & @xmath00.30 & @xmath00.24 & & @xmath301.48 & @xmath301.43 & @xmath300.20 & @xmath00.03 & @xmath300.75 + 3711c7 & @xmath00.56 & @xmath00.00 & @xmath00.27 & @xmath00.22 & @xmath00.09 & @xmath00.42 & @xmath00.23 & @xmath300.47 & @xmath300.50 & @xmath00.24 & @xmath00.06 & @xmath00.04 + 3733c3 & @xmath00.47 & @xmath300.01 & @xmath00.22 & @xmath00.14 & @xmath00.15 & @xmath00.22 & @xmath300.04 & @xmath300.18 & @xmath300.17 & @xmath00.16 & @xmath00.03 & @xmath00.35 + 3796c6 & @xmath00.75 & @xmath300.08 & @xmath00.37 & @xmath00.19 & @xmath00.27 & @xmath00.24 & @xmath300.04 & @xmath300.83 & @xmath300.81 & @xmath00.13 & @xmath00.09 & @xmath00.25 + 3965c6 & @xmath00.44 & @xmath00.18 & @xmath00.28 & @xmath00.18 & @xmath00.21 & @xmath00.28 & @xmath00.02 & @xmath300.08 & @xmath300.09 & @xmath00.24 & @xmath00.18 & @xmath00.59 + 4085c3 & @xmath00.48 & @xmath300.02 & @xmath00.44 & @xmath00.29 & @xmath00.32 & @xmath00.39 & @xmath00.10 & @xmath300.59 & @xmath300.57 & @xmath00.14 & @xmath00.06 & @xmath00.11 + 4217c6 & @xmath00.70 & @xmath00.08 & @xmath00.37 & @xmath00.37 & @xmath00.42 & @xmath00.36 & @xmath00.05 & @xmath300.57 & @xmath300.57 & @xmath00.27 & @xmath00.10 & @xmath00.44 + 4263c6 & @xmath00.57 & @xmath00.22 & @xmath00.32 & @xmath00.35 & @xmath00.24 & @xmath00.31 & @xmath00.12 & @xmath300.33 & @xmath300.34 & @xmath00.30 & @xmath00.09 & @xmath00.45 + 431c2 & @xmath00.16 & @xmath00.14 & @xmath00.28 & @xmath00.17 & @xmath00.09 & @xmath00.17 & @xmath00.02 & @xmath300.02 & @xmath300.04 & @xmath00.21 & @xmath00.13 & @xmath00.33 + 4365c3 & & @xmath300.09 & @xmath00.10 & @xmath00.05 & @xmath00.18 & @xmath00.31 & @xmath300.24 & @xmath300.80 & @xmath300.79 & @xmath300.18 & @xmath00.00 & @xmath300.33 + 4478c8 & @xmath00.34 & @xmath300.07 & @xmath00.36 & @xmath00.25 & @xmath00.24 & @xmath00.32 & @xmath00.03 & @xmath300.35 & @xmath300.36 & @xmath00.11 & @xmath00.08 & @xmath300.01 + 455c1 & @xmath00.46 & @xmath300.05 & @xmath00.32 & @xmath00.20 & @xmath00.33 & @xmath00.42 & @xmath300.06 & @xmath300.62 & @xmath300.60 & @xmath300.03 & @xmath00.03 & @xmath300.17 + 4612c6 & @xmath00.48 & @xmath300.21 & @xmath00.25 & @xmath00.29 & @xmath00.33 & @xmath00.38 & @xmath00.11 & @xmath300.79 & @xmath300.76 & @xmath300.03 & @xmath00.06 & @xmath300.16 + 4740c8 & @xmath00.53 & @xmath300.08 & @xmath00.33 & @xmath00.23 & @xmath00.28 & @xmath00.37 & @xmath300.04 & @xmath300.49 & @xmath300.49 & @xmath00.21 & @xmath00.01 & @xmath300.10 + 4876c6 & @xmath00.53 & @xmath00.04 & @xmath00.34 & @xmath00.24 & @xmath00.30 & @xmath00.32 & @xmath00.00 & @xmath300.59 & @xmath300.59 & @xmath00.09 & @xmath00.09 & @xmath00.11 + 5319c6 & @xmath00.56 & @xmath300.15 & @xmath00.25 & @xmath00.17 & @xmath00.35 & @xmath00.33 & @xmath300.01 & @xmath300.40 & @xmath300.40 & @xmath00.13 & @xmath00.07 & @xmath00.07 + 5351c8 & @xmath00.48 & @xmath300.07 & @xmath00.22 & @xmath00.10 & @xmath00.20 & @xmath00.07 & @xmath300.08 & @xmath300.09 & @xmath300.08 & @xmath00.19 & @xmath00.08 & @xmath00.36 + 5400c8 & & @xmath00.01 & @xmath00.06 & @xmath300.07 & @xmath00.11 & @xmath00.03 & @xmath300.07 & @xmath00.13 & @xmath00.10 & @xmath00.07 & @xmath00.14 & @xmath00.17 + 5487c8 & & @xmath300.09 & @xmath00.22 & @xmath00.11 & @xmath00.14 & @xmath00.16 & @xmath300.09 & @xmath300.47 & @xmath300.47 & @xmath00.03 & @xmath00.03 & @xmath00.04 + 5543c6 & @xmath300.34 & @xmath00.03 & @xmath00.17 & @xmath00.09 & @xmath00.26 & @xmath00.19 & @xmath00.06 & @xmath300.17 & @xmath300.16 & @xmath300.18 & @xmath00.03 & @xmath300.36 + 5588c6 & @xmath00.44 & @xmath300.07 & @xmath00.33 & @xmath00.05 & @xmath00.20 & @xmath00.19 & @xmath300.03 & @xmath300.20 & @xmath300.19 & @xmath00.15 & @xmath00.04 & @xmath00.09 + 5664c6 & @xmath300.09 & @xmath00.27 & @xmath00.00 & @xmath00.12 & @xmath00.14 & @xmath00.29 & @xmath00.10 & @xmath00.23 & & @xmath00.03 & @xmath00.08 & @xmath00.24 + 5908c6 & @xmath00.51 & @xmath300.07 & @xmath00.32 & @xmath00.22 & @xmath00.25 & @xmath00.27 & @xmath300.08 & @xmath300.67 & @xmath300.67 & @xmath00.10 & @xmath00.06 & @xmath300.01 + 5977c6 & @xmath300.16 & @xmath00.28 & @xmath00.15 & @xmath00.05 & @xmath00.26 & @xmath00.06 & @xmath300.12 & @xmath00.17 & @xmath00.20 & @xmath00.01 & @xmath00.13 & @xmath00.25 + 5980c6 & & @xmath300.10 & @xmath00.30 & @xmath00.11 & @xmath00.18 & @xmath00.18 & @xmath00.06 & @xmath300.17 & @xmath300.16 & @xmath00.25 & @xmath00.17 & @xmath00.69 + 608c1 & @xmath00.68 & @xmath300.31 & & & @xmath00.31 & @xmath00.30 & @xmath00.06 & @xmath301.74 & @xmath301.74 & & @xmath300.08 & @xmath300.84 + 6090c6 & @xmath300.34 & @xmath300.11 & @xmath00.04 & @xmath00.02 & @xmath00.20 & @xmath00.03 & @xmath300.05 & @xmath300.02 & & @xmath300.17 & @xmath00.06 & @xmath300.15 + 6164c6 & @xmath00.20 & @xmath300.07 & @xmath00.21 & @xmath00.14 & @xmath00.09 & @xmath00.20 & @xmath300.11 & @xmath300.11 & @xmath300.14 & @xmath00.07 & @xmath00.10 & @xmath00.22 + 6230c5 & & @xmath300.01 & @xmath00.01 & @xmath300.13 & @xmath300.01 & @xmath300.01 & @xmath300.06 & @xmath00.19 & @xmath00.21 & @xmath300.05 & @xmath00.02 & @xmath00.28 + 6263c6 & & @xmath00.20 & @xmath300.08 & @xmath300.14 & @xmath00.13 & @xmath00.00 & @xmath300.22 & @xmath00.48 & & @xmath00.07 & @xmath00.26 & @xmath00.29 + 6391c8 & @xmath00.48 & @xmath300.12 & @xmath00.24 & @xmath00.27 & @xmath00.25 & @xmath00.32 & @xmath300.10 & @xmath300.64 & @xmath300.61 & @xmath00.07 & @xmath00.06 & @xmath300.07 + 6419c5 & @xmath00.33 & @xmath300.01 & @xmath00.24 & @xmath00.24 & @xmath00.18 & @xmath00.23 & @xmath00.06 & @xmath300.24 & @xmath300.24 & @xmath00.25 & @xmath00.13 & @xmath00.51 + 6426c8 & & @xmath00.04 & @xmath00.41 & @xmath00.21 & @xmath00.35 & @xmath00.35 & @xmath00.06 & @xmath300.39 & @xmath300.41 & @xmath00.06 & @xmath00.11 & @xmath00.21 + 6505c6 & @xmath00.76 & @xmath300.16 & @xmath00.32 & @xmath00.19 & @xmath00.22 & @xmath00.23 & @xmath300.07 & @xmath300.37 & @xmath300.38 & @xmath00.12 & @xmath00.06 & @xmath00.09 + 650c2 & @xmath00.40 & @xmath00.01 & @xmath00.39 & @xmath00.26 & @xmath00.29 & @xmath00.36 & @xmath300.04 & @xmath300.66 & @xmath300.67 & @xmath00.09 & @xmath00.04 & @xmath00.13 + 6549c6 & @xmath00.10 & @xmath300.01 & @xmath300.02 & @xmath00.06 & @xmath300.02 & @xmath00.04 & @xmath300.02 & @xmath00.34 & @xmath00.32 & @xmath00.09 & @xmath00.11 & @xmath00.33 + 6637c8 & @xmath00.55 & @xmath00.07 & @xmath00.26 & @xmath00.14 & @xmath00.28 & @xmath00.30 & @xmath300.04 & @xmath300.26 & @xmath300.26 & @xmath00.11 & @xmath00.14 & @xmath00.16 + 6717c6 & @xmath00.62 & @xmath300.02 & @xmath00.24 & @xmath00.18 & @xmath00.12 & @xmath00.12 & @xmath300.02 & @xmath300.18 & @xmath300.19 & @xmath00.08 & @xmath00.09 & @xmath00.25 + 6828c7 & @xmath300.38 & @xmath300.02 & @xmath00.05 & @xmath300.01 & @xmath00.11 & @xmath00.06 & @xmath300.12 & @xmath00.12 & @xmath00.08 & @xmath300.03 & @xmath00.07 & @xmath00.15 + 6913c7 & @xmath00.32 & @xmath300.07 & @xmath00.26 & @xmath00.18 & @xmath00.23 & @xmath00.22 & @xmath300.04 & @xmath300.28 & @xmath300.26 & @xmath00.18 & @xmath00.10 & @xmath00.20 + 867c3 & @xmath00.58 & @xmath300.07 & @xmath00.32 & @xmath00.19 & @xmath00.24 & @xmath00.26 & @xmath300.17 & @xmath300.47 & @xmath300.47 & @xmath00.12 & @xmath00.06 & @xmath00.12 + + 119799c4 & 0.19 & 0.08 & 0.03 & 0.08 & 0.10 & 0.14 & 0.09 & 0.11 & 0.27 & 0.05 & 0.05 & 0.15 + 129499c4 & & 0.06 & 0.04 & 0.05 & 0.09 & 0.09 & 0.05 & 0.10 & & 0.05 & 0.04 & + 176772c5 & 0.20 & 0.12 & 0.04 & 0.10 & 0.08 & 0.17 & 0.09 & 0.13 & 0.25 & 0.04 & 0.03 & + 181349c5 & 0.14 & 0.10 & 0.03 & 0.07 & 0.08 & 0.12 & 0.07 & 0.11 & 0.21 & 0.05 & 0.04 & + 183783c5 & & 0.09 & 0.04 & 0.07 & 0.08 & 0.12 & 0.07 & 0.10 & & 0.05 & 0.04 & + 184088c5 & 0.17 & 0.09 & 0.03 & 0.07 & 0.09 & 0.14 & 0.09 & 0.10 & 0.27 & 0.05 & 0.05 & 0.15 + 184618c5 & & 0.06 & 0.03 & 0.05 & 0.08 & 0.10 & 0.07 & 0.10 & 0.20 & 0.06 & 0.04 & 0.15 + 185169c5 & & 0.05 & 0.03 & 0.05 & 0.09 & 0.11 & 0.06 & 0.10 & & 0.04 & 0.04 & 0.16 + 185357c5 & 0.19 & 0.10 & 0.04 & 0.08 & 0.09 & 0.14 & 0.08 & 0.12 & 0.25 & 0.04 & 0.04 & + 185541c5 & 0.16 & 0.10 & 0.03 & 0.07 & 0.08 & 0.14 & 0.08 & 0.10 & 0.23 & 0.05 & 0.04 & + 187067c5 & & 0.03 & 0.04 & 0.03 & 0.09 & 0.06 & 0.03 & 0.09 & 0.17 & 0.06 & 0.03 & 0.05 + 193190c5 & & 0.10 & 0.04 & 0.07 & 0.09 & 0.13 & 0.07 & 0.11 & 0.22 & 0.03 & 0.04 & + 197366c5 & 0.19 & 0.13 & 0.03 & 0.08 & 0.09 & 0.17 & 0.08 & 0.12 & 0.27 & 0.05 & 0.04 & + 215681c6 & 0.15 & 0.07 & 0.04 & 0.06 & 0.09 & 0.11 & 0.06 & 0.11 & & 0.04 & 0.04 & 0.16 + 216922c6 & 0.20 & 0.10 & 0.03 & 0.08 & 0.09 & 0.15 & 0.09 & 0.10 & 0.26 & 0.05 & 0.05 & + 218198c6 & 0.22 & 0.13 & 0.04 & 0.10 & 0.09 & 0.18 & 0.11 & 0.12 & 0.29 & 0.04 & 0.04 & + 219909c6 & 0.12 & 0.05 & 0.04 & 0.04 & 0.09 & 0.09 & 0.05 & 0.09 & 0.19 & 0.04 & 0.04 & + 221537c6 & 0.14 & 0.08 & 0.04 & 0.06 & 0.09 & 0.12 & 0.07 & 0.12 & & 0.05 & 0.04 & + 223113c6 & 0.13 & 0.04 & 0.03 & 0.04 & 0.09 & 0.09 & 0.05 & 0.10 & 0.23 & 0.06 & 0.06 & + 223343c6 & 0.18 & 0.09 & 0.04 & 0.07 & 0.09 & 0.14 & 0.08 & 0.12 & 0.24 & 0.04 & 0.04 & + 223621c6 & & 0.06 & 0.04 & 0.05 & 0.09 & 0.12 & 0.04 & 0.09 & 0.18 & 0.05 & 0.04 & + 223722c6 & 0.13 & 0.05 & 0.04 & 0.04 & 0.10 & 0.07 & 0.05 & 0.11 & 0.20 & 0.05 & 0.03 & + 224206c6 & & 0.08 & 0.04 & 0.06 & 0.09 & 0.13 & 0.07 & 0.12 & 0.29 & 0.05 & 0.05 & + 224866c6 & 0.20 & 0.11 & 0.03 & 0.09 & 0.09 & 0.16 & 0.10 & 0.12 & 0.26 & 0.05 & 0.05 & + 224951c6 & 0.15 & 0.09 & 0.04 & 0.07 & 0.08 & 0.13 & 0.07 & 0.12 & & 0.04 & 0.04 & + 225531c6 & 0.16 & 0.07 & 0.03 & 0.06 & 0.09 & 0.11 & 0.07 & 0.09 & 0.20 & 0.05 & 0.04 & 0.15 + 226450c6 & 0.14 & 0.10 & 0.04 & 0.09 & 0.09 & 0.17 & 0.10 & 0.14 & & 0.04 & 0.04 & + 226850c6 & & 0.03 & 0.03 & 0.03 & 0.08 & 0.05 & 0.03 & 0.09 & 0.16 & 0.06 & 0.05 & 0.05 + 227867c6 & 0.13 & 0.07 & 0.04 & 0.06 & 0.08 & 0.12 & 0.07 & 0.10 & 0.20 & 0.05 & 0.04 & 0.15 + 228466c6 & 0.10 & 0.03 & 0.04 & 0.03 & 0.09 & 0.05 & 0.03 & 0.10 & 0.17 & 0.05 & 0.04 & 0.07 + 229507c6 & & 0.05 & 0.04 & 0.05 & 0.10 & 0.08 & 0.06 & 0.12 & 0.20 & 0.05 & 0.04 & 0.15 + 230424c6 & 0.17 & 0.08 & 0.03 & 0.07 & 0.09 & 0.13 & 0.08 & 0.09 & 0.23 & 0.04 & 0.04 & 0.14 + 230483c6 & 0.16 & 0.09 & 0.04 & 0.07 & 0.08 & 0.13 & 0.09 & 0.10 & 0.23 & 0.05 & 0.04 & + 231379c6 & 0.16 & 0.09 & 0.03 & 0.08 & 0.09 & 0.14 & 0.09 & 0.10 & 0.22 & 0.05 & 0.04 & 0.16 + 231618c6 & 0.18 & 0.09 & 0.04 & 0.08 & 0.09 & 0.14 & 0.09 & 0.12 & 0.25 & 0.05 & 0.04 & 0.16 + 232493c6 & 0.12 & 0.03 & 0.03 & 0.03 & 0.09 & 0.07 & 0.05 & 0.09 & 0.21 & 0.05 & 0.04 & 0.05 + 233121c6 & 0.15 & 0.07 & 0.03 & 0.07 & 0.09 & 0.10 & 0.09 & 0.10 & 0.23 & 0.06 & 0.05 & + 233560c6 & 0.19 & 0.11 & 0.04 & 0.09 & 0.09 & 0.16 & 0.10 & 0.11 & 0.26 & 0.04 & 0.04 & 0.16 + 233708c6 & 0.20 & 0.10 & 0.04 & 0.08 & 0.10 & 0.16 & 0.09 & 0.14 & 0.30 & 0.04 & 0.05 & 0.15 + 240059c6 & 0.15 & 0.07 & 0.05 & 0.06 & 0.11 & 0.11 & 0.07 & 0.10 & 0.22 & 0.04 & 0.04 & + 240083c6 & 0.16 & 0.08 & 0.04 & 0.07 & 0.09 & 0.11 & 0.09 & 0.11 & 0.22 & 0.06 & 0.04 & + 259050c7 & 0.18 & 0.10 & 0.04 & 0.09 & 0.10 & 0.14 & 0.10 & 0.11 & 0.26 & 0.04 & 0.04 & + 259377c7 & & 0.06 & 0.04 & 0.05 & 0.08 & 0.10 & 0.05 & 0.09 & & 0.05 & 0.04 & 0.16 + 260308c7 & 0.18 & 0.11 & 0.04 & 0.09 & 0.09 & 0.15 & 0.08 & 0.13 & 0.30 & 0.04 & 0.04 & + 262018c7 & 0.14 & 0.09 & 0.04 & 0.08 & 0.09 & 0.11 & 0.07 & 0.11 & 0.21 & 0.04 & 0.04 & 0.15 + 266442c7 & 0.20 & 0.10 & 0.04 & 0.09 & 0.09 & 0.16 & 0.10 & 0.12 & 0.29 & 0.04 & 0.04 & 0.18 + 270316c7 & & 0.09 & 0.04 & 0.07 & 0.10 & 0.15 & 0.09 & 0.11 & 0.26 & 0.04 & 0.04 & 0.15 + 275181c7 & 0.17 & 0.09 & 0.04 & 0.07 & 0.09 & 0.13 & 0.08 & 0.10 & 0.21 & 0.04 & 0.04 & 0.16 + 277490c7 & 0.19 & 0.11 & 0.04 & 0.09 & 0.09 & 0.16 & 0.12 & 0.12 & 0.27 & 0.05 & 0.03 & 0.14 + 278419c7 & 0.15 & 0.08 & 0.04 & 0.06 & 0.08 & 0.12 & 0.07 & 0.11 & & 0.05 & 0.04 & + 282804c7 & 0.14 & 0.09 & 0.04 & 0.07 & 0.09 & 0.15 & 0.08 & 0.13 & & 0.03 & 0.04 & 0.16 + 286252c7 & & 0.10 & 0.04 & 0.07 & 0.08 & 0.16 & 0.07 & 0.11 & 0.24 & 0.05 & 0.04 & 0.14 + 45512c2 & 0.16 & 0.09 & 0.04 & 0.08 & 0.08 & 0.12 & 0.09 & 0.11 & 0.22 & 0.06 & 0.05 & 0.17 + 47188c2 & 0.15 & 0.09 & 0.04 & 0.08 & 0.09 & 0.13 & 0.08 & 0.10 & & 0.04 & 0.04 & + 77186c3 & 0.16 & 0.09 & 0.03 & 0.07 & 0.08 & 0.13 & 0.07 & 0.10 & 0.21 & 0.05 & 0.04 & 0.16 + 77707c3 & 0.16 & 0.08 & 0.04 & 0.07 & 0.08 & 0.12 & 0.07 & 0.11 & 0.21 & 0.05 & 0.04 & 0.16 + 80582c3 & & 0.06 & 0.03 & 0.05 & 0.08 & 0.11 & 0.05 & 0.09 & 0.22 & 0.05 & 0.04 & 0.15 + 81644c3 & 0.19 & 0.10 & 0.04 & 0.08 & 0.09 & 0.16 & 0.08 & 0.12 & 0.26 & 0.05 & 0.04 & 0.15 + 82227c3 & 0.15 & 0.07 & 0.04 & 0.06 & 0.09 & 0.12 & 0.06 & 0.10 & & 0.05 & 0.04 & 0.15 + 83531c3 & 0.17 & 0.11 & 0.03 & 0.07 & 0.08 & 0.15 & 0.07 & 0.10 & 0.21 & 0.05 & 0.03 & 0.14 + 84255c3 & 0.17 & 0.08 & 0.05 & 0.07 & 0.10 & 0.13 & 0.09 & 0.12 & 0.27 & 0.04 & 0.04 & 0.14 + 86757c3 & & 0.10 & 0.04 & 0.08 & 0.09 & 0.13 & 0.09 & 0.12 & 0.22 & 0.04 & 0.03 & 0.16 + 88522c3 & & 0.09 & 0.03 & 0.07 & 0.09 & 0.15 & 0.08 & 0.12 & 0.29 & 0.05 & 0.04 & 0.15 + + 225847c6 & & 0.09 & 0.04 & 0.09 & 0.08 & 0.13 & 0.10 & 0.12 & 0.22 & 0.07 & 0.05 & + 227379c6 & 0.19 & 0.12 & 0.04 & 0.09 & 0.09 & 0.16 & 0.09 & 0.13 & 0.24 & 0.04 & 0.03 & + 228407c6 & & 0.10 & 0.04 & 0.08 & 0.10 & 0.15 & 0.09 & 0.13 & 0.28 & 0.03 & 0.04 & 0.17 + 230208c6 & & 0.06 & 0.04 & 0.05 & 0.09 & 0.11 & 0.06 & 0.10 & 0.24 & 0.04 & 0.04 & 0.15 + 239284c6 & & 0.11 & 0.04 & 0.09 & 0.10 & 0.18 & 0.10 & 0.13 & & 0.03 & 0.05 & + 265795c7 & & 0.09 & 0.04 & 0.08 & 0.09 & 0.15 & 0.09 & 0.12 & & 0.05 & 0.05 & 0.17 + 268360c7 & & 0.09 & 0.04 & 0.07 & 0.08 & 0.14 & 0.07 & 0.09 & 0.21 & 0.05 & 0.04 & 0.16 + 268493c7 & & 0.10 & 0.03 & 0.09 & 0.08 & 0.14 & 0.10 & 0.11 & 0.23 & 0.06 & 0.03 & 0.15 + 271021c7 & & 0.07 & 0.04 & 0.06 & 0.09 & 0.11 & 0.07 & 0.12 & 0.26 & 0.03 & 0.04 & 0.16 + 271400c7 & & 0.10 & 0.04 & 0.09 & 0.08 & 0.16 & 0.09 & 0.11 & 0.29 & 0.04 & 0.04 & 0.16 + 77182c3 & & 0.07 & 0.04 & 0.06 & 0.09 & 0.11 & 0.08 & 0.11 & 0.26 & 0.04 & 0.05 & + 85597c3 & & 0.11 & 0.04 & 0.09 & 0.09 & 0.16 & 0.09 & 0.12 & & 0.04 & 0.04 & 0.17 + + 1156c2 & 0.19 & 0.11 & 0.04 & 0.09 & 0.09 & 0.14 & 0.10 & 0.12 & 0.24 & 0.05 & 0.04 & 0.16 + 1407c3 & 0.10 & 0.03 & 0.04 & 0.03 & 0.08 & 0.04 & 0.03 & 0.09 & 0.18 & 0.05 & 0.04 & 0.05 + 1491c7 & 0.13 & 0.05 & 0.04 & 0.05 & 0.09 & 0.09 & 0.06 & 0.10 & & 0.04 & 0.04 & 0.13 + 1554c7 & 0.14 & 0.03 & 0.04 & 0.03 & 0.06 & 0.06 & 0.02 & 0.10 & & 0.04 & 0.02 & 0.07 + 166c3 & 0.10 & 0.03 & 0.04 & 0.03 & 0.07 & 0.06 & 0.04 & 0.09 & 0.16 & 0.06 & 0.04 & 0.06 + 1754c3 & & 0.03 & 0.04 & 0.04 & 0.10 & 0.08 & 0.05 & 0.10 & 0.23 & 0.04 & 0.05 & 0.12 + 1814c1 & 0.16 & 0.07 & 0.03 & 0.06 & 0.09 & 0.12 & 0.07 & 0.10 & 0.22 & 0.06 & 0.05 & 0.16 + 1876c2 & & 0.03 & 0.04 & 0.04 & 0.07 & 0.05 & 0.03 & 0.10 & 0.16 & 0.03 & 0.03 & 0.04 + 1917c1 & & 0.09 & 0.04 & 0.07 & 0.08 & 0.12 & 0.07 & 0.11 & 0.21 & 0.05 & 0.04 & 0.17 + 1918c1 & 0.11 & 0.04 & 0.03 & 0.03 & 0.08 & 0.06 & 0.04 & 0.10 & 0.20 & 0.06 & 0.04 & 0.09 + 201583c3 & & 0.03 & 0.04 & 0.03 & 0.08 & 0.04 & & 0.10 & 0.17 & & 0.04 & 0.07 + 2110c7 & 0.18 & 0.09 & 0.04 & 0.07 & 0.08 & 0.16 & 0.08 & 0.11 & 0.29 & 0.05 & 0.05 & 0.17 + 2178c7 & 0.15 & 0.08 & 0.04 & 0.07 & 0.09 & 0.12 & 0.07 & 0.11 & 0.22 & 0.04 & 0.04 & 0.12 + 2200c3 & & 0.05 & 0.04 & 0.04 & 0.08 & 0.08 & 0.04 & 0.10 & 0.17 & 0.04 & 0.04 & 0.13 + 2220c7 & 0.12 & 0.04 & 0.04 & 0.03 & 0.08 & 0.08 & 0.02 & 0.09 & & 0.04 & 0.03 & 0.13 + 222c3 & & 0.03 & 0.03 & 0.03 & 0.08 & 0.06 & 0.03 & 0.09 & 0.18 & 0.05 & 0.04 & 0.10 + 2335c2 & 0.09 & 0.04 & 0.05 & 0.04 & 0.09 & 0.06 & 0.03 & 0.11 & 0.18 & 0.04 & 0.04 & 0.04 + 2407c2 & 0.09 & 0.02 & 0.03 & 0.03 & 0.07 & 0.05 & 0.03 & 0.09 & 0.17 & 0.06 & 0.04 & 0.06 + 2422c7 & 0.15 & 0.09 & 0.04 & 0.08 & 0.09 & 0.14 & 0.09 & 0.13 & & 0.05 & 0.04 & 0.15 + 2470c3 & 0.14 & 0.08 & 0.04 & 0.06 & 0.09 & 0.13 & 0.07 & 0.11 & 0.23 & 0.04 & 0.04 & 0.15 + 2502c3 & 0.10 & 0.04 & 0.04 & 0.04 & 0.09 & 0.07 & 0.04 & 0.10 & 0.18 & 0.04 & 0.04 & 0.08 + 2532c6 & 0.14 & 0.06 & 0.04 & 0.05 & 0.10 & 0.10 & 0.06 & 0.11 & 0.25 & 0.04 & 0.05 & 0.17 + 2580c6 & 0.19 & 0.11 & 0.03 & 0.07 & 0.09 & 0.15 & 0.08 & 0.11 & 0.26 & 0.05 & 0.04 & 0.15 + 2769c3 & 0.14 & 0.07 & 0.04 & 0.06 & 0.08 & 0.11 & 0.06 & 0.10 & 0.21 & 0.05 & 0.04 & 0.14 + 2772c7 & 0.12 & 0.04 & 0.03 & 0.04 & 0.08 & 0.08 & 0.05 & 0.09 & 0.19 & 0.06 & 0.05 & 0.14 + 2812c8 & 0.16 & 0.07 & 0.04 & 0.07 & 0.09 & 0.12 & 0.09 & 0.10 & 0.26 & 0.05 & 0.05 & 0.15 + 2947c3 & 0.09 & 0.02 & 0.03 & 0.03 & 0.07 & 0.04 & 0.03 & 0.09 & 0.17 & 0.06 & 0.04 & 0.08 + 2948c7 & 0.14 & 0.12 & 0.04 & 0.09 & 0.10 & 0.18 & 0.10 & 0.16 & & 0.04 & 0.04 & 0.15 + 3018c3 & 0.14 & 0.07 & 0.04 & 0.06 & 0.09 & 0.11 & 0.07 & 0.09 & 0.23 & 0.04 & 0.04 & 0.14 + 3035c7 & 0.10 & 0.03 & 0.04 & 0.03 & 0.08 & 0.05 & 0.03 & 0.10 & 0.19 & 0.05 & 0.04 & 0.07 + 3091c8 & 0.12 & 0.04 & 0.03 & 0.04 & 0.09 & 0.09 & 0.05 & 0.08 & 0.22 & 0.06 & 0.05 & 0.11 + 3101c7 & 0.10 & 0.04 & 0.03 & 0.04 & 0.07 & 0.07 & 0.04 & 0.10 & 0.18 & 0.06 & 0.03 & 0.09 + 3142c3 & 0.13 & 0.04 & 0.03 & 0.04 & 0.08 & 0.07 & 0.05 & 0.09 & 0.19 & 0.06 & 0.05 & 0.12 + 3161c3 & & 0.04 & 0.04 & 0.04 & 0.08 & 0.08 & 0.03 & 0.09 & 0.17 & 0.05 & 0.04 & 0.12 + 3191c7 & 0.11 & 0.03 & 0.03 & 0.03 & 0.09 & 0.06 & 0.04 & 0.10 & 0.22 & 0.05 & 0.05 & 0.11 + 3201c6 & 0.11 & 0.04 & 0.04 & 0.03 & 0.08 & 0.10 & 0.04 & 0.09 & 0.22 & 0.05 & 0.05 & 0.13 + 3238c6 & & 0.05 & 0.04 & 0.05 & 0.09 & 0.10 & 0.06 & 0.10 & 0.24 & 0.06 & 0.05 & 0.14 + 3267c3 & 0.16 & 0.07 & 0.04 & 0.06 & 0.09 & 0.11 & 0.07 & 0.11 & 0.22 & 0.05 & 0.04 & 0.16 + 3515c3 & & 0.09 & 0.04 & 0.07 & 0.08 & 0.11 & 0.07 & 0.10 & 0.19 & 0.04 & 0.04 & 0.14 + 3558c6 & & 0.06 & 0.03 & 0.05 & 0.08 & 0.10 & 0.06 & 0.10 & 0.20 & 0.06 & 0.04 & 0.15 + 3690c7 & 0.08 & 0.03 & 0.05 & 0.04 & 0.06 & 0.03 & & 0.10 & 0.15 & 0.06 & 0.02 & 0.06 + 3711c7 & 0.09 & 0.02 & 0.03 & 0.02 & 0.07 & 0.06 & 0.03 & 0.07 & 0.15 & 0.06 & 0.04 & 0.06 + 3733c3 & 0.13 & 0.06 & 0.03 & 0.05 & 0.09 & 0.10 & 0.07 & 0.10 & 0.21 & 0.06 & 0.05 & 0.15 + 3796c6 & 0.13 & 0.06 & 0.04 & 0.06 & 0.10 & 0.09 & 0.06 & 0.10 & 0.21 & 0.04 & 0.04 & 0.15 + 3965c6 & 0.19 & 0.11 & 0.04 & 0.09 & 0.09 & 0.17 & 0.09 & 0.12 & 0.26 & 0.04 & 0.04 & 0.17 + 4085c3 & 0.13 & 0.07 & 0.04 & 0.06 & 0.10 & 0.10 & 0.07 & 0.10 & 0.23 & 0.04 & 0.04 & 0.14 + 4217c6 & 0.13 & 0.07 & 0.04 & 0.06 & 0.10 & 0.11 & 0.07 & 0.10 & 0.20 & 0.03 & 0.04 & 0.16 + 4263c6 & 0.11 & 0.03 & 0.03 & 0.03 & 0.08 & 0.08 & 0.04 & 0.09 & 0.19 & 0.06 & 0.04 & 0.12 + 431c2 & 0.12 & 0.06 & 0.04 & 0.04 & 0.08 & 0.12 & 0.04 & 0.08 & 0.21 & 0.06 & 0.04 & 0.13 + 4365c3 & & 0.03 & 0.04 & 0.03 & 0.06 & 0.04 & 0.03 & 0.10 & 0.16 & 0.05 & 0.03 & 0.05 + 4478c8 & 0.12 & 0.03 & 0.03 & 0.04 & 0.08 & 0.06 & 0.05 & 0.10 & 0.19 & 0.06 & 0.04 & 0.09 + 455c1 & 0.07 & 0.02 & 0.04 & 0.03 & 0.08 & 0.07 & 0.02 & 0.10 & 0.15 & 0.04 & 0.03 & 0.05 + 4612c6 & 0.10 & 0.03 & 0.03 & 0.03 & 0.08 & 0.06 & 0.03 & 0.09 & 0.17 & 0.05 & 0.04 & 0.06 + 4740c8 & 0.08 & 0.03 & 0.04 & 0.03 & 0.08 & 0.07 & 0.03 & 0.10 & 0.17 & 0.04 & 0.03 & 0.08 + 4876c6 & 0.10 & 0.04 & 0.04 & 0.04 & 0.09 & 0.09 & 0.04 & 0.09 & 0.17 & 0.05 & 0.04 & 0.11 + 5319c6 & 0.16 & 0.07 & 0.04 & 0.07 & 0.09 & 0.13 & 0.08 & 0.10 & 0.26 & 0.04 & 0.04 & 0.13 + 5351c8 & 0.14 & 0.07 & 0.04 & 0.07 & 0.09 & 0.10 & 0.07 & 0.10 & 0.21 & 0.05 & 0.04 & 0.16 + 5400c8 & & 0.08 & 0.03 & 0.05 & 0.08 & 0.12 & 0.05 & 0.08 & 0.19 & 0.06 & 0.04 & 0.12 + 5487c8 & & 0.05 & 0.04 & 0.04 & 0.09 & 0.08 & 0.05 & 0.10 & 0.19 & 0.04 & 0.03 & 0.11 + 5543c6 & 0.19 & 0.08 & 0.04 & 0.08 & 0.09 & 0.12 & 0.09 & 0.12 & 0.26 & 0.06 & 0.04 & 0.08 + 5588c6 & 0.13 & 0.04 & 0.03 & 0.04 & 0.09 & 0.08 & 0.05 & 0.10 & 0.21 & 0.06 & 0.04 & 0.11 + 5664c6 & 0.14 & 0.08 & 0.04 & 0.06 & 0.08 & 0.12 & 0.06 & 0.11 & & 0.05 & 0.04 & 0.16 + 5908c6 & 0.09 & 0.03 & 0.04 & 0.03 & 0.07 & 0.04 & 0.03 & 0.09 & 0.19 & 0.05 & 0.02 & 0.04 + 5977c6 & 0.18 & 0.08 & 0.04 & 0.07 & 0.09 & 0.12 & 0.08 & 0.13 & 0.28 & 0.05 & 0.05 & 0.17 + 5980c6 & & 0.05 & 0.03 & 0.05 & 0.09 & 0.12 & 0.06 & 0.09 & 0.22 & 0.07 & 0.06 & 0.14 + 608c1 & 0.06 & 0.04 & & & 0.07 & 0.03 & 0.02 & 0.09 & 0.13 & & 0.02 & 0.02 + 6090c6 & 0.15 & 0.10 & 0.04 & 0.09 & 0.08 & 0.13 & 0.10 & 0.12 & & 0.05 & 0.03 & 0.11 + 6164c6 & 0.14 & 0.06 & 0.03 & 0.06 & 0.08 & 0.11 & 0.07 & 0.10 & 0.20 & 0.07 & 0.04 & 0.13 + 6230c5 & & 0.07 & 0.04 & 0.05 & 0.08 & 0.11 & 0.06 & 0.10 & 0.21 & 0.06 & 0.04 & 0.13 + 6263c6 & & 0.13 & 0.04 & 0.08 & 0.08 & 0.18 & 0.09 & 0.14 & & 0.05 & 0.03 & 0.15 + 6391c8 & 0.10 & 0.03 & 0.04 & 0.03 & 0.07 & 0.04 & 0.03 & 0.11 & 0.15 & 0.05 & 0.04 & 0.07 + 6419c5 & 0.13 & 0.07 & 0.04 & 0.06 & 0.09 & 0.11 & 0.07 & 0.10 & 0.21 & 0.03 & 0.04 & 0.18 + 6426c8 & & 0.04 & 0.03 & 0.03 & 0.09 & 0.09 & 0.04 & 0.09 & 0.21 & 0.04 & 0.04 & 0.12 + 6505c6 & 0.09 & 0.03 & 0.03 & 0.03 & 0.07 & 0.05 & 0.04 & 0.10 & 0.15 & 0.06 & 0.04 & 0.10 + 650c2 & 0.18 & 0.09 & 0.04 & 0.08 & 0.10 & 0.12 & 0.10 & 0.11 & 0.26 & 0.06 & 0.05 & 0.16 + 6549c6 & 0.15 & 0.08 & 0.04 & 0.06 & 0.09 & 0.13 & 0.06 & 0.10 & 0.23 & 0.05 & 0.04 & 0.15 + 6637c8 & 0.13 & 0.04 & 0.03 & 0.05 & 0.10 & 0.09 & 0.06 & 0.10 & 0.24 & 0.04 & 0.05 & 0.13 + 6717c6 & 0.13 & 0.05 & 0.03 & 0.05 & 0.09 & 0.11 & 0.06 & 0.09 & 0.20 & 0.07 & 0.05 & 0.13 + 6828c7 & 0.18 & 0.08 & 0.04 & 0.07 & 0.10 & 0.13 & 0.08 & 0.12 & 0.28 & 0.04 & 0.04 & 0.16 + 6913c7 & 0.09 & 0.04 & 0.05 & 0.03 & 0.09 & 0.08 & 0.04 & 0.10 & 0.17 & 0.03 & 0.04 & 0.11 + 867c3 & 0.08 & 0.03 & 0.04 & 0.03 & 0.08 & 0.06 & 0.03 & 0.10 & 0.18 & 0.04 & 0.04 & 0.09 + + 119799c4 & 0.08 & 0.02 & 0.00 & 0.02 & 0.12 & 0.04 & 0.17 & 0.13 & 0.08 & 0.09 & 0.10 & 0.08 + 129499c4 & & 0.04 & 0.07 & 0.11 & 0.17 & 0.12 & 0.10 & 0.14 & & 0.04 & 0.18 & + 176772c5 & 0.08 & 0.01 & 0.04 & 0.04 & 0.15 & 0.15 & 0.08 & 0.17 & 0.08 & 0.05 & 0.16 & + 181349c5 & 0.08 & 0.25 & 0.07 & 0.08 & 0.21 & 0.08 & 0.18 & 0.15 & 0.08 & 0.07 & 0.16 & + 183783c5 & & & 0.10 & 0.08 & 0.06 & 0.11 & 0.22 & 0.13 & & 0.08 & 0.18 & + 184088c5 & 0.08 & 0.00 & 0.07 & 0.00 & 0.11 & 0.09 & 0.16 & 0.13 & 0.00 & 0.04 & 0.12 & 0.08 + 184618c5 & & 0.11 & 0.07 & 0.00 & 0.15 & 0.10 & 0.13 & 0.14 & 0.08 & 0.08 & 0.10 & 0.08 + 185169c5 & & 0.07 & 0.07 & 0.00 & 0.08 & 0.05 & 0.16 & 0.14 & & 0.08 & 0.10 & 0.08 + 185357c5 & 0.08 & 0.04 & 0.00 & 0.07 & 0.12 & 0.05 & 0.10 & 0.13 & 0.08 & 0.02 & 0.14 & + 185541c5 & 0.08 & 0.07 & 0.02 & 0.07 & 0.21 & 0.09 & 0.13 & 0.15 & 0.01 & 0.08 & 0.13 & + 187067c5 & & 0.07 & 0.08 & 0.08 & 0.09 & 0.10 & 0.08 & 0.15 & 0.02 & 0.08 & 0.15 & 0.08 + 193190c5 & & 0.07 & 0.04 & 0.08 & 0.17 & 0.19 & 0.09 & 0.12 & 0.08 & 0.14 & 0.09 & + 197366c5 & 0.08 & 0.06 & 0.08 & 0.08 & 0.10 & 0.13 & 0.12 & 0.15 & 0.08 & 0.05 & 0.13 & + 215681c6 & 0.08 & 0.01 & 0.00 & 0.04 & 0.12 & 0.11 & 0.12 & 0.13 & & 0.05 & 0.15 & 0.08 + 216922c6 & 0.08 & 0.05 & 0.00 & 0.00 & 0.11 & 0.06 & 0.05 & 0.14 & 0.08 & 0.11 & 0.12 & + 218198c6 & 0.08 & 0.01 & 0.00 & 0.04 & 0.05 & 0.09 & 0.04 & 0.14 & 0.08 & 0.08 & 0.16 & + 219909c6 & 0.08 & 0.06 & 0.00 & 0.00 & 0.12 & 0.08 & 0.14 & 0.14 & 0.08 & 0.08 & 0.13 & + 221537c6 & 0.08 & 0.00 & 0.00 & 0.08 & 0.15 & 0.09 & 0.13 & 0.13 & & 0.00 & 0.15 & + 223113c6 & 0.08 & 0.12 & 0.00 & 0.08 & 0.07 & 0.09 & 0.08 & 0.13 & 0.11 & 0.06 & 0.09 & + 223343c6 & 0.08 & 0.00 & 0.08 & 0.00 & 0.11 & 0.15 & 0.14 & 0.13 & 0.05 & 0.11 & 0.12 & + 223621c6 & & 0.06 & 0.00 & 0.11 & 0.07 & 0.12 & 0.09 & 0.14 & 0.08 & 0.07 & 0.12 & + 223722c6 & 0.08 & 0.08 & 0.04 & 0.04 & 0.10 & 0.08 & 0.06 & 0.14 & 0.11 & 0.09 & 0.14 & + 224206c6 & & 0.05 & 0.06 & 0.05 & 0.15 & 0.16 & 0.10 & 0.14 & 0.08 & 0.08 & 0.15 & + 224866c6 & 0.08 & 0.07 & 0.11 & 0.01 & 0.11 & 0.06 & 0.10 & 0.14 & 0.08 & 0.04 & 0.12 & + 224951c6 & 0.08 & 0.08 & 0.07 & 0.04 & 0.05 & 0.15 & 0.20 & 0.13 & & 0.16 & 0.15 & + 225531c6 & 0.08 & 0.04 & 0.08 & & 0.12 & 0.08 & 0.07 & 0.13 & 0.09 & 0.08 & 0.11 & 0.08 + 226450c6 & 0.08 & 0.03 & 0.07 & 0.04 & 0.15 & 0.02 & 0.12 & 0.15 & & 0.00 & 0.11 & + 226850c6 & & 0.07 & 0.00 & 0.04 & 0.13 & 0.11 & 0.07 & 0.12 & 0.10 & 0.04 & 0.10 & 0.08 + 227867c6 & 0.08 & 0.18 & 0.04 & 0.01 & 0.11 & 0.10 & 0.14 & 0.14 & 0.04 & 0.01 & 0.12 & 0.08 + 228466c6 & 0.08 & 0.11 & 0.00 & 0.04 & 0.09 & 0.10 & 0.19 & 0.15 & 0.18 & 0.04 & 0.14 & 0.08 + 229507c6 & & 0.01 & 0.07 & 0.14 & 0.11 & 0.08 & 0.14 & 0.12 & 0.08 & 0.04 & 0.17 & 0.08 + 230424c6 & 0.08 & 0.01 & 0.00 & 0.14 & 0.06 & 0.12 & 0.11 & 0.12 & 0.06 & 0.09 & 0.13 & 0.08 + 230483c6 & 0.08 & 0.00 & 0.14 & 0.08 & 0.15 & 0.07 & 0.14 & 0.14 & 0.08 & 0.04 & 0.10 & + 231379c6 & 0.08 & 0.07 & 0.02 & 0.11 & 0.11 & 0.09 & 0.13 & 0.14 & 0.04 & 0.08 & 0.08 & 0.08 + 231618c6 & 0.08 & 0.04 & 0.04 & 0.00 & 0.11 & 0.13 & 0.17 & 0.12 & 0.08 & 0.04 & 0.14 & 0.08 + 232493c6 & 0.08 & 0.07 & & 0.07 & 0.10 & 0.12 & 0.11 & 0.14 & 0.04 & 0.08 & 0.16 & 0.08 + 233121c6 & 0.08 & 0.02 & 0.08 & 0.04 & 0.07 & 0.08 & 0.06 & 0.12 & 0.05 & 0.07 & 0.09 & + 233560c6 & 0.08 & 0.07 & 0.00 & 0.05 & 0.05 & 0.09 & 0.12 & 0.12 & 0.06 & 0.04 & 0.11 & 0.08 + 233708c6 & 0.08 & 0.03 & 0.07 & 0.05 & 0.14 & 0.11 & 0.11 & 0.12 & 0.10 & 0.05 & 0.11 & 0.08 + 240059c6 & 0.08 & 0.01 & 0.07 & 0.04 & 0.11 & 0.09 & 0.10 & 0.11 & 0.08 & 0.02 & 0.07 & + 240083c6 & 0.08 & 0.11 & 0.01 & 0.04 & 0.10 & 0.08 & 0.07 & 0.11 & 0.08 & 0.04 & 0.13 & + 259050c7 & 0.08 & 0.16 & 0.08 & 0.04 & 0.17 & 0.03 & 0.18 & 0.15 & 0.08 & 0.08 & 0.16 & + 259377c7 & & 0.04 & 0.01 & 0.11 & 0.16 & 0.13 & 0.12 & 0.14 & & 0.06 & 0.11 & 0.08 + 260308c7 & 0.08 & 0.01 & 0.08 & 0.02 & 0.15 & 0.12 & 0.07 & 0.14 & 0.08 & 0.09 & 0.15 & + 262018c7 & 0.08 & 0.02 & 0.07 & 0.02 & 0.16 & 0.10 & 0.14 & 0.15 & 0.08 & 0.08 & 0.13 & 0.08 + 266442c7 & 0.08 & 0.04 & 0.08 & 0.02 & 0.10 & 0.11 & 0.12 & 0.13 & 0.08 & 0.00 & 0.12 & 0.08 + 270316c7 & & 0.04 & 0.08 & 0.08 & 0.08 & 0.10 & 0.10 & 0.14 & 0.08 & 0.10 & 0.11 & 0.08 + 275181c7 & 0.08 & 0.11 & 0.08 & 0.00 & 0.13 & 0.13 & 0.08 & 0.14 & 0.11 & 0.07 & 0.12 & 0.08 + 277490c7 & 0.08 & 0.02 & 0.00 & 0.01 & 0.09 & 0.07 & 0.15 & 0.13 & 0.03 & 0.01 & 0.11 & 0.08 + 278419c7 & 0.08 & 0.04 & 0.00 & 0.13 & 0.13 & 0.14 & 0.08 & 0.13 & & 0.08 & 0.15 & + 282804c7 & 0.08 & 0.01 & 0.08 & 0.04 & 0.17 & 0.17 & 0.08 & 0.15 & & 0.01 & 0.16 & 0.08 + 286252c7 & & 0.03 & 0.08 & 0.08 & 0.15 & 0.11 & 0.06 & 0.12 & 0.02 & 0.13 & 0.16 & 0.08 + 45512c2 & 0.08 & 0.11 & 0.08 & 0.00 & 0.15 & 0.09 & 0.08 & 0.14 & 0.09 & 0.01 & 0.12 & 0.08 + 47188c2 & 0.08 & 0.14 & 0.08 & 0.08 & 0.11 & 0.08 & 0.19 & 0.15 & & 0.08 & 0.12 & + 77186c3 & 0.08 & 0.13 & 0.04 & 0.04 & 0.18 & 0.10 & 0.09 & 0.13 & 0.08 & 0.12 & 0.15 & 0.08 + 77707c3 & 0.08 & 0.02 & 0.00 & 0.05 & 0.12 & 0.05 & 0.15 & 0.14 & 0.07 & 0.08 & 0.15 & 0.08 + 80582c3 & & 0.06 & 0.04 & 0.08 & 0.11 & 0.05 & 0.11 & 0.13 & 0.08 & 0.01 & 0.13 & 0.08 + 81644c3 & 0.08 & 0.00 & 0.00 & 0.07 & 0.17 & 0.04 & 0.17 & 0.14 & 0.08 & 0.04 & 0.14 & 0.08 + 82227c3 & 0.08 & 0.21 & 0.04 & 0.13 & 0.15 & 0.12 & 0.01 & 0.12 & & 0.01 & 0.16 & 0.08 + 83531c3 & 0.08 & 0.01 & 0.02 & 0.08 & 0.10 & 0.15 & 0.18 & 0.14 & 0.08 & 0.04 & 0.13 & 0.08 + 84255c3 & 0.08 & 0.01 & 0.02 & 0.06 & 0.05 & 0.12 & 0.17 & 0.10 & 0.04 & 0.08 & 0.15 & 0.08 + 86757c3 & & 0.08 & 0.05 & 0.00 & 0.09 & 0.10 & 0.11 & 0.13 & 0.08 & 0.06 & 0.11 & 0.08 + 88522c3 & & 0.16 & 0.04 & 0.11 & 0.12 & 0.17 & 0.08 & 0.13 & 0.08 & 0.01 & 0.13 & 0.08 + + 225847c6 & & 0.08 & 0.06 & 0.06 & 0.07 & 0.13 & 0.12 & 0.14 & 0.02 & 0.09 & 0.11 & + 227379c6 & 0.08 & 0.03 & 0.04 & 0.06 & 0.10 & 0.13 & 0.13 & 0.12 & 0.12 & 0.00 & 0.15 & + 228407c6 & & 0.01 & 0.04 & 0.02 & 0.16 & 0.12 & 0.13 & 0.15 & 0.08 & 0.04 & 0.11 & 0.08 + 230208c6 & & 0.00 & 0.09 & 0.01 & 0.12 & 0.10 & 0.14 & 0.13 & 0.04 & 0.08 & 0.07 & 0.08 + 239284c6 & & 0.04 & 0.08 & 0.04 & 0.16 & 0.12 & 0.11 & 0.14 & & 0.01 & 0.14 & + 265795c7 & & 0.08 & 0.08 & 0.06 & 0.18 & 0.03 & 0.16 & 0.16 & & 0.05 & 0.14 & 0.08 + 268360c7 & & 0.06 & 0.04 & 0.06 & 0.11 & 0.00 & 0.14 & 0.11 & 0.08 & 0.00 & 0.13 & 0.08 + 268493c7 & & 0.08 & 0.07 & 0.09 & 0.16 & 0.04 & 0.18 & 0.16 & 0.13 & 0.06 & 0.10 & 0.08 + 271021c7 & & 0.01 & 0.11 & 0.06 & 0.11 & 0.11 & 0.14 & 0.15 & 0.08 & 0.11 & 0.16 & 0.08 + 271400c7 & & 0.02 & 0.06 & 0.02 & 0.09 & 0.09 & 0.11 & 0.14 & 0.08 & 0.08 & 0.15 & 0.08 + 77182c3 & & 0.01 & 0.00 & 0.04 & 0.13 & 0.12 & 0.16 & 0.12 & 0.08 & 0.05 & 0.07 & + 85597c3 & & 0.04 & 0.07 & 0.02 & 0.12 & 0.11 & 0.10 & 0.13 & & 0.14 & 0.15 & 0.08 + + 1156c2 & 0.08 & 0.02 & 0.00 & 0.01 & 0.04 & 0.07 & 0.08 & 0.10 & 0.08 & 0.05 & 0.11 & 0.08 + 1407c3 & 0.08 & 0.02 & 0.07 & 0.03 & 0.10 & 0.14 & 0.22 & 0.12 & 0.11 & 0.08 & 0.11 & 0.08 + 1491c7 & 0.08 & 0.03 & 0.00 & 0.00 & 0.13 & 0.12 & 0.06 & 0.16 & & 0.18 & 0.13 & 0.08 + 1554c7 & 0.08 & 0.04 & 0.07 & 0.00 & 0.13 & 0.05 & 0.09 & 0.14 & & 0.11 & 0.12 & 0.08 + 166c3 & 0.08 & 0.14 & 0.08 & 0.08 & 0.08 & 0.05 & 0.13 & 0.14 & 0.08 & 0.18 & 0.15 & 0.08 + 1754c3 & & 0.06 & 0.02 & 0.00 & 0.10 & 0.13 & 0.10 & 0.12 & 0.04 & 0.11 & 0.10 & 0.08 + 1814c1 & 0.08 & 0.07 & 0.04 & 0.08 & 0.10 & 0.11 & 0.11 & 0.12 & 0.01 & 0.04 & 0.09 & 0.08 + 1876c2 & & 0.00 & 0.04 & 0.07 & 0.12 & 0.14 & 0.11 & 0.14 & 0.08 & 0.04 & 0.13 & 0.08 + 1917c1 & & 0.08 & 0.11 & 0.11 & 0.15 & 0.14 & 0.16 & 0.13 & 0.08 & 0.11 & 0.15 & 0.08 + 1918c1 & 0.08 & 0.11 & 0.00 & 0.08 & 0.12 & 0.10 & 0.11 & 0.13 & 0.08 & 0.04 & 0.09 & 0.08 + 201583c3 & & 0.14 & 0.08 & 0.08 & 0.13 & 0.16 & & 0.16 & 0.08 & & 0.09 & 0.08 + 2110c7 & 0.08 & 0.01 & 0.04 & 0.09 & 0.09 & 0.10 & 0.12 & 0.12 & 0.05 & 0.01 & 0.12 & 0.08 + 2178c7 & 0.08 & 0.04 & 0.04 & 0.04 & 0.08 & 0.05 & 0.05 & 0.12 & 0.08 & 0.00 & 0.10 & 0.08 + 2200c3 & & 0.11 & 0.06 & 0.07 & 0.06 & 0.10 & 0.17 & 0.14 & 0.09 & 0.08 & 0.07 & 0.08 + 2220c7 & 0.08 & 0.03 & 0.05 & 0.02 & 0.12 & 0.10 & 0.09 & 0.14 & & 0.07 & 0.13 & 0.08 + 222c3 & & 0.04 & 0.02 & 0.07 & 0.09 & 0.02 & 0.03 & 0.12 & 0.03 & 0.00 & 0.08 & 0.08 + 2335c2 & 0.08 & 0.04 & 0.04 & 0.07 & 0.07 & 0.08 & 0.02 & 0.11 & 0.08 & 0.08 & 0.11 & 0.08 + 2407c2 & 0.08 & 0.05 & 0.07 & 0.06 & 0.08 & 0.06 & 0.14 & 0.11 & 0.12 & 0.08 & 0.08 & 0.08 + 2422c7 & 0.08 & 0.11 & 0.04 & 0.08 & 0.09 & 0.07 & 0.20 & 0.12 & & 0.00 & 0.08 & 0.08 + 2470c3 & 0.08 & 0.04 & 0.00 & 0.06 & 0.05 & 0.08 & 0.16 & 0.12 & 0.12 & 0.00 & 0.08 & 0.08 + 2502c3 & 0.08 & 0.02 & 0.08 & 0.00 & 0.06 & 0.09 & 0.02 & 0.13 & 0.08 & 0.03 & 0.11 & 0.08 + 2532c6 & 0.08 & 0.05 & 0.11 & 0.04 & 0.15 & 0.07 & 0.16 & 0.13 & 0.08 & 0.07 & 0.12 & 0.08 + 2580c6 & 0.08 & 0.11 & 0.06 & 0.00 & 0.15 & 0.10 & 0.12 & 0.18 & 0.08 & 0.01 & 0.14 & 0.08 + 2769c3 & 0.08 & 0.11 & 0.04 & 0.07 & 0.07 & 0.07 & 0.11 & 0.13 & 0.08 & 0.08 & 0.13 & 0.08 + 2772c7 & 0.08 & 0.01 & 0.07 & 0.00 & 0.09 & 0.08 & 0.10 & 0.11 & 0.04 & 0.04 & 0.08 & 0.08 + 2812c8 & 0.08 & 0.01 & 0.01 & 0.08 & 0.07 & 0.12 & 0.07 & 0.11 & 0.00 & 0.04 & 0.11 & 0.08 + 2947c3 & 0.08 & 0.04 & 0.14 & 0.04 & 0.12 & 0.09 & 0.09 & 0.13 & 0.08 & 0.11 & 0.08 & 0.08 + 2948c7 & 0.08 & 0.00 & 0.07 & 0.04 & 0.09 & 0.06 & 0.21 & 0.13 & & 0.05 & 0.18 & 0.08 + 3018c3 & 0.08 & 0.04 & 0.00 & 0.01 & 0.08 & 0.05 & 0.10 & 0.11 & 0.06 & 0.08 & 0.09 & 0.08 + 3035c7 & 0.08 & 0.04 & 0.04 & 0.00 & 0.11 & 0.07 & 0.12 & 0.13 & 0.14 & 0.07 & 0.08 & 0.08 + 3091c8 & 0.08 & 0.01 & 0.02 & 0.04 & 0.10 & 0.08 & 0.04 & 0.12 & 0.06 & 0.08 & 0.07 & 0.08 + 3101c7 & 0.08 & 0.05 & 0.04 & 0.02 & 0.10 & 0.05 & 0.08 & 0.13 & 0.03 & 0.06 & 0.09 & 0.08 + 3142c3 & 0.08 & 0.09 & 0.07 & 0.05 & 0.07 & 0.05 & 0.06 & 0.13 & 0.11 & 0.01 & 0.10 & 0.08 + 3161c3 & & 0.02 & 0.00 & 0.05 & 0.11 & 0.07 & 0.12 & 0.12 & 0.12 & 0.04 & 0.10 & 0.08 + 3191c7 & 0.08 & 0.05 & 0.01 & 0.00 & 0.10 & 0.05 & 0.07 & 0.11 & 0.01 & 0.00 & 0.08 & 0.08 + 3201c6 & 0.08 & 0.04 & 0.00 & 0.09 & 0.10 & 0.07 & 0.12 & 0.13 & 0.18 & 0.11 & 0.12 & 0.08 + 3238c6 & & 0.04 & 0.04 & 0.09 & 0.13 & 0.04 & 0.10 & 0.12 & 0.00 & 0.01 & 0.10 & 0.08 + 3267c3 & 0.08 & 0.06 & 0.04 & 0.01 & 0.06 & 0.10 & 0.11 & 0.09 & 0.05 & 0.01 & 0.08 & 0.08 + 3515c3 & & 0.03 & 0.00 & 0.05 & 0.12 & 0.09 & 0.09 & 0.12 & 0.08 & 0.01 & 0.09 & 0.08 + 3558c6 & & 0.01 & 0.11 & 0.03 & 0.06 & 0.09 & 0.15 & 0.13 & 0.08 & 0.07 & 0.10 & 0.08 + 3690c7 & 0.08 & 0.08 & 0.00 & 0.08 & 0.11 & 0.09 & & 0.12 & 0.06 & 0.08 & 0.13 & 0.08 + 3711c7 & 0.08 & 0.04 & 0.08 & 0.07 & 0.08 & 0.06 & 0.20 & 0.13 & 0.11 & 0.14 & 0.13 & 0.08 + 3733c3 & 0.08 & 0.01 & 0.02 & 0.01 & 0.10 & 0.08 & 0.13 & 0.11 & 0.05 & 0.01 & 0.10 & 0.08 + 3796c6 & 0.08 & 0.01 & 0.05 & 0.02 & 0.12 & 0.04 & 0.11 & 0.12 & 0.06 & 0.04 & 0.09 & 0.08 + 3965c6 & 0.08 & 0.03 & 0.00 & 0.00 & 0.03 & 0.12 & 0.09 & 0.12 & 0.08 & 0.01 & 0.11 & 0.08 + 4085c3 & 0.08 & 0.02 & 0.07 & 0.00 & 0.08 & 0.09 & 0.08 & 0.10 & 0.04 & 0.07 & 0.10 & 0.08 + 4217c6 & 0.08 & 0.01 & 0.08 & 0.08 & 0.12 & 0.08 & 0.02 & 0.15 & 0.02 & 0.04 & 0.14 & 0.08 + 4263c6 & 0.08 & 0.06 & 0.08 & 0.04 & 0.12 & 0.07 & 0.18 & 0.14 & 0.11 & 0.08 & 0.11 & 0.08 + 431c2 & 0.08 & 0.03 & 0.00 & 0.07 & 0.10 & 0.11 & 0.12 & 0.12 & 0.03 & 0.00 & 0.10 & 0.08 + 4365c3 & & 0.20 & 0.08 & 0.07 & 0.12 & 0.17 & 0.19 & 0.14 & 0.17 & 0.08 & 0.16 & 0.08 + 4478c8 & 0.08 & 0.05 & 0.01 & 0.00 & 0.10 & 0.08 & 0.04 & 0.10 & 0.09 & 0.02 & 0.10 & 0.08 + 455c1 & 0.08 & 0.05 & 0.00 & 0.08 & 0.09 & 0.12 & 0.08 & 0.11 & 0.08 & 0.08 & 0.11 & 0.08 + 4612c6 & 0.08 & 0.08 & 0.06 & 0.08 & 0.14 & 0.11 & 0.12 & 0.13 & 0.08 & 0.05 & 0.06 & 0.08 + 4740c8 & 0.08 & 0.00 & 0.01 & 0.02 & 0.15 & 0.08 & 0.13 & 0.11 & 0.08 & 0.10 & 0.10 & 0.08 + 4876c6 & 0.08 & 0.08 & 0.00 & 0.00 & 0.08 & 0.09 & 0.06 & 0.11 & 0.05 & 0.08 & 0.13 & 0.08 + 5319c6 & 0.08 & 0.14 & 0.04 & 0.01 & 0.05 & 0.07 & 0.18 & 0.12 & 0.10 & 0.08 & 0.10 & 0.08 + 5351c8 & 0.08 & 0.06 & 0.04 & 0.05 & 0.07 & 0.07 & 0.14 & 0.11 & 0.06 & 0.04 & 0.10 & 0.08 + 5400c8 & & 0.06 & 0.04 & 0.02 & 0.07 & 0.07 & 0.08 & 0.12 & 0.08 & 0.04 & 0.09 & 0.08 + 5487c8 & & 0.01 & 0.04 & 0.02 & 0.11 & 0.06 & 0.16 & 0.14 & 0.15 & 0.08 & 0.10 & 0.08 + 5543c6 & 0.08 & 0.06 & 0.07 & 0.00 & 0.16 & 0.10 & 0.07 & 0.13 & 0.08 & 0.12 & 0.17 & 0.08 + 5588c6 & 0.08 & 0.09 & 0.04 & 0.00 & 0.04 & 0.05 & 0.11 & 0.11 & 0.06 & 0.03 & 0.10 & 0.08 + 5664c6 & 0.08 & 0.00 & 0.08 & 0.00 & 0.20 & 0.14 & 0.17 & 0.13 & & 0.01 & 0.11 & 0.08 + 5908c6 & 0.08 & 0.12 & 0.00 & 0.07 & 0.07 & 0.11 & 0.12 & 0.11 & 0.10 & 0.08 & 0.08 & 0.08 + 5977c6 & 0.08 & 0.00 & 0.04 & 0.04 & 0.12 & 0.10 & 0.15 & 0.12 & 0.01 & 0.05 & 0.11 & 0.08 + 5980c6 & & 0.09 & 0.04 & 0.06 & 0.08 & 0.04 & 0.06 & 0.11 & 0.08 & 0.04 & 0.12 & 0.08 + 608c1 & 0.08 & 0.08 & & & 0.16 & 0.15 & 0.13 & 0.14 & 0.06 & & 0.12 & 0.08 + 6090c6 & 0.08 & 0.11 & 0.14 & 0.00 & 0.15 & 0.14 & 0.19 & 0.12 & & 0.08 & 0.12 & 0.08 + 6164c6 & 0.08 & 0.01 & 0.00 & 0.08 & 0.11 & 0.10 & 0.12 & 0.11 & 0.11 & 0.05 & 0.09 & 0.08 + 6230c5 & & 0.04 & 0.00 & 0.01 & 0.05 & 0.06 & 0.09 & 0.12 & 0.05 & 0.03 & 0.10 & 0.08 + 6263c6 & & 0.11 & 0.07 & 0.02 & 0.08 & 0.19 & 0.16 & 0.14 & & 0.11 & 0.16 & 0.08 + 6391c8 & 0.08 & 0.13 & 0.07 & 0.04 & 0.08 & 0.09 & 0.11 & 0.12 & 0.08 & 0.11 & 0.06 & 0.08 + 6419c5 & 0.08 & 0.12 & 0.00 & 0.02 & 0.07 & 0.05 & 0.09 & 0.13 & 0.18 & 0.01 & 0.14 & 0.08 + 6426c8 & & 0.00 & 0.00 & 0.02 & 0.15 & 0.12 & 0.07 & 0.12 & 0.02 & 0.08 & 0.09 & 0.08 + 6505c6 & 0.08 & 0.04 & 0.00 & 0.05 & 0.05 & 0.07 & 0.08 & 0.11 & 0.14 & 0.04 & 0.07 & 0.08 + 650c2 & 0.08 & 0.00 & 0.04 & 0.08 & 0.07 & 0.06 & 0.06 & 0.08 & 0.06 & 0.02 & 0.07 & 0.08 + 6549c6 & 0.08 & 0.09 & 0.04 & 0.07 & 0.11 & 0.02 & 0.17 & 0.13 & 0.06 & 0.06 & 0.10 & 0.08 + 6637c8 & 0.08 & 0.01 & 0.00 & 0.08 & 0.11 & 0.04 & 0.12 & 0.11 & 0.08 & 0.00 & 0.12 & 0.08 + 6717c6 & 0.08 & 0.01 & 0.01 & 0.00 & 0.07 & 0.09 & 0.10 & 0.11 & 0.08 & 0.07 & 0.09 & 0.08 + 6828c7 & 0.08 & 0.01 & 0.00 & 0.05 & 0.11 & 0.10 & 0.15 & 0.10 & 0.04 & 0.04 & 0.11 & 0.08 + 6913c7 & 0.08 & 0.00 & 0.04 & 0.07 & 0.09 & 0.05 & 0.08 & 0.11 & 0.08 & 0.04 & 0.07 & 0.08 + 867c3 & 0.08 & 0.02 & 0.00 & 0.08 & 0.11 & 0.08 & 0.10 & 0.08 & 0.06 & 0.08 & 0.08 & 0.08 + | we present radial velocities and chemical abundances of o , na , mg , al , si , ca , cr , fe , co , ni , and cu for a sample of 156 red giant branch stars in two galactic bulge fields centered near ( l , b)=(@xmath05.25,3.02 ) and ( 0,12 ) .
the ( @xmath05.25,3.02 ) field also includes observations of the bulge globular cluster ngc 6553 .
the results are based on high resolution ( r@xmath120,000 ) , high signal to noise ( s / n@xmath270 ) flames giraffe spectra obtained through the eso archive . however , we only selected a subset of the original observations that included spectra with both high s / n and that did not show strong tio absorption bands .
the present work extends previous analyses of this data set beyond fe and the @xmath3elements mg , si , ca , and ti .
while we find reasonable agreement with past work , the data presented here indicate that the bulge may exhibit a different chemical composition than the local thick disk , especially at [ fe / h]@xmath20.5
. in particular , the bulge [ @xmath3/fe ] ratios may remain enhanced to a slightly higher [ fe / h ] than the thick disk and the fe
peak elements co , ni , and cu appear enhanced compared to the disk .
there is also some evidence that the [ na / fe ] ( but not [ al / fe ] ) trends between the bulge and local disk may be different at low and high metallicity .
we also find that the velocity dispersion decreases as a function of increasing [ fe / h ] for both fields , and do not detect any significant cold , high velocity population .
a comparison with chemical enrichment models indicates that a significant fraction of hypernovae are required to explain the bulge abundance trends , and that initial mass functions that are steep , top heavy ( and do not include strong outflow ) , or truncated to avoid including contributions from stars @xmath440 m@xmath5 are ruled out , in particular because of disagreement with the fe
peak abundance data . for most elements ,
the ngc 6553 stars exhibit nearly identical abundance trends to comparable metallicity bulge field stars .
however , the star to star scatter and mean [ na / fe ] ratios appear higher in the cluster , perhaps indicating additional self enrichment . |
a new concept , all - optical particle radiation detector has been very recently proposed based on the mechanism of upconversion in re - doped materials @xcite .
this process , in which low energy incident optical radiation ( infrared light - ir ) is converted into higher energy emitted photons ( visible light ) , is efficiently accomplished by incorporating re ions in inorganic matrices due to their @xmath5-electrons configurations .
in fact , ground state absorption ( level 0 ) allows the rare earth ions to reach a metastable intermediate state ( level 1 ) , characterized by relatively long lifetimes ( @xmath6ms ) , then another photon delivered by a pump laser tuned to the transition @xmath7 promotes the ion to a more energetic state ( level 2 ) .
radiative transition from this latter excited state back to the ground state is then observed by means of traditional detectors as photomultipliers ( pmt ) or photodiodes ( pd ) . to date , this mechanism has been extensively applied for the development of lasers and optical devices @xcite but its applicability in the field of particle detection has not been deeply investigated . the visible light yield ( ly ) of this novel device
depends on the upconversion efficiency and on the number of re - ions excited into the metastable level 1 per energy unit of the particle .
this latter quantity can be estimated by studying the ly in the ir band which very few articles in literature are concerned with @xcite . actually , there is little interest in the re - doped crystals infrared scintillation for their long decay lifetimes and for the low quantum efficiency of pmt in this spectral region .
the aim of this work is to propose a method that allows us to tackle a systematic investigation of the ir ly in several different materials , composed of different matrices , dopants and concentrations .
this method is based on the luminescence comparison with a reference pr:(lu@xmath0y@xmath1)@xmath2al@xmath3o@xmath4 single crystal whose light yield in the near uv - visible is known .
moreover , this method is applied to this crystal , thereby yielding its ir ly .
the mixed lutetium - yttrium aluminum garnet sample has been grown via czochralsky method at itme , warsaw and its preparation is described elsewhere @xcite .
the interest in mixed pr : luyag crystals is related to its much better performance in terms of light yield and energy resolution as compared to either pr : luag or pr : yag . the @xmath8 mm@xmath9 sample chosen for the present measurements has a reported light yield of and a energy resolution in the uv - vis range @xcite . to investigate its light yield in the ir band , the pr : luyag sample is excited by x - rays generated by an electron gun that can be operated both in continuous and in pulsed mode sweeping the electron beam at frequency @xcite .
the several intense electron beam impinges on a @xmath10-thick tantalum foil placed in front of the sample to make sure that the whole crystal is excited only by x - rays .
the scintillation response of the crystal sample is measured using si ( mod .
hamamtsu s1337 - 1010bq ) and ingaas ( mod .
thorlabs det20c ) pds .
the small contribution to the pd signal of the x - rays that are not absorbed in the crystal and reach the pd can be estimated and then subtracted by covering the pd with an thin aluminum foil . in order to verify that the radioluminescence is determined by the pr@xmath11 ions emission ,
the x - rays excitation spectra are compared with those obtained when the crystal is irradiated with a pulsed , frequency - quadrupled nd : yag laser ( ) . in fact , whereas the host matrix is transparent at this wavelength , the pr@xmath11 ions are directly excited into @xmath12 levels , thus allowing us to simulate the energy transfer process from the electron - hole recombination to the re ions after the x - ray excitation .
we have recorded spectra of the pr : luyag emission due to x - ray and uv excitation ( fig.[fig : egun ] ) . the oceanoptics
650 redtide and oceanoptics nirquest512 spectrometers were used for the 200 and 900 regions , respectively . in both spectra
one can clearly see emissions related to the same pr@xmath11 energy levels .
we observe a strong @xmath13 emission in the range from 300 to and narrow lines in the visible / near - infrared region due to emission from levels @xmath14 , @xmath15 , @xmath16 .
the spectra below are similar to those reported by previous authors in a pr : luag crystal @xcite . the main near - infrared transitions that we identify are : @xmath17 @xmath18 , @xmath19 @xmath20 ( 900 ) and @xmath17 @xmath21 ( 1400 ) .
the emission from @xmath21 is expected to be partially non radiatively quenched , whereas the lower lying levels are strongly quenched in luyag matrix . as a consequence
no mid - infrared emission is expected from these latter levels .
we measure the total number of charge carriers generated per x - ray pulse in the photodiode @xmath22 with @xmath23 being the time integrated photodiode signal , @xmath24 the impedance of oscilloscope , @xmath25 and @xmath26 which are the gain and the time constant of the active integrator , respectively .
if the pd quantum efficiency @xmath27 is constant in the considered range , @xmath28 is also given by equation @xmath29 where @xmath30 is the energy deposited in the sample , @xmath31 the crystal reflectivity , @xmath32 is the solid angle subtended by the pd with area @xmath33 located at a distance @xmath34 from the crystal .
moreover , the measured charge per pulse @xmath35 is related through a proportionality constant @xmath36 to the energy released by x - ray pulse in the crystal .
in fact , as shown in the fig.[linearity ] , the measured luminescence @xmath23 linearly depends on @xmath35 .
it is then possible to obtain a general expression for the measured light yield in a definite wavelength range : @xmath37 as in the point source approximation the following expression holds true @xmath38 the parameters @xmath39 and @xmath40 can be obtained by a fit of data recorded at different relative distances @xmath41 of the photodiode from the scintillating crystal , as shown in fig.[poso ] .
we observe that in this type of measurements it crucial to precisely know the efficiency of the x - ray generation process , related to the previously mentioned proportionality constant @xmath36 .
the latter can be estimated if the x - ray energy is fully released in the sample and provided the ly and the @xmath42 parameter of a generic crystal in any wavelength range are known . as the sensitivity of the detector that has been used to measure the light yield of our luyag : pr crystal reported in @xcite peaks in the uv range , it is reasonable to expect that the reported light yield is mainly determined by @xmath13 radiative transitions .
in addition , due to fast integration time the slow components might have been underestimated . in order to select the photodiode signal component due to the uv scintillation
, the previously described measurements are repeated at a fixed distance with optical longpass filters ( thorlabs , fgl and fel filter sets ) located in front of the photodiodes .
the use of filters allows us to estimate the light yield in narrower bandwidths and with a better accuracy by taking into account the wavelength dependence of the photodiode s responsivity . as the energy of the x - rays
can be assumed to be in the range of few tens of kev , the ly of our sample measured in @xcite at has to be reduced by 10% , in agreement with previously published data @xcite .
we report in table [ tab : res ] the results of the light yield measurements in several optical ranges .
.measured light yield of the luyag : pr crystal for different optical ranges . [ cols="<,^",options="header " , ] [ tab : res ]
we have demonstrated a practical way to accurately estimate the infrared light yield of a re - doped crystal by comparison with the near uv - visible luminescence of a reference crystal .
the presented method allows us to make accurate ly measurements since it is based on the point - source approximation , as verified by performing measurements at several distances from the source , and because it is possible to vary the energy released per pulse in the crystal and to average over several measurements .
furthermore , the use of several optical longpass filters reduces the error due to the wavelength dependence of the photodiode quantum efficiency .
the method has been applied to a ( lu@xmath0y@xmath1)@xmath2al@xmath3o@xmath4 crystal whose emission in the uv range had been previously measured . in spite of its high ly in the uv ,
its emission in the near infrared band is limited to a few thousands of ph / mev , probably due to the @xmath43 manifold quenching .
although the states of our interest @xmath44 and @xmath45 , characterized by ms - long radiative lifetimes , are efficiently populated by the relaxation of the higher manifolds , no mid - infrared emission is expected from them in this host matrix .
the data reported in this work can be used to extend the ly measurements up to the mid - infrared band , provided that photodiodes with a lower band gap than ingaas are used . in order to identify the most suitable crystals for the development of the upconversion - based detector ,
several re - doped crystals are currently being investigated with the method described in this work .
the preliminary results obtained for nd : yag and tm : yag are particularly encouraging for our aims and will soon be published .
the growth of the crystals used in this research has been financed from the funds of the polish national science centre granted on the basis of decision no .
dec-2012/05/b / st5/00324 . | we propose a new technique to measure the infrared scintillation light yield of rare earth ( re ) doped crystals by comparing it to near uv - visible scintillation of a calibrated pr:(lu@xmath0y@xmath1)@xmath2al@xmath3o@xmath4 sample . as an example
, we apply this technique to provide the light yield in visible and infrared range up to of this crystal
. infrared light yield , pr : luyag , radioluminescence |
semilattices are ordered structures that admit an algebraic presentation .
formally , a semilattice may be presented as a partially ordered set @xmath0 such that for every pair @xmath1 there exists the supremum or _ join _
@xmath2 of the set @xmath3 .
equivalently , @xmath4 is a semilattice if @xmath5 is an idempotent commutative semigroup operation . in this case
we call @xmath6 a _ @xmath5-semilattice _
( `` join - semilattice '' ) .
we can also define a partial _ meet _ operation @xmath7 in a @xmath5-semilattice that equals the infimum of @xmath3 whenever it exists .
the paradigmatic example of a semilattice is given by any family of sets that is closed under ( finite ) union .
actually , the free semilattice on @xmath8 generators is given by the powerset of @xmath9 minus @xmath10 .
a reference for concepts of general algebra and ideals in semilattices is @xcite .
the aim of this paper is to obtain an inner characterization of direct product decompositions of semilattices akin to those in classical algebra . to attain this goal we represent these decompositions by means of factor congruences .
a _ factor congruence _ is the kernel @xmath11 of a projection @xmath12 onto a direct factor of @xmath6 .
thus , a direct product representation @xmath13 is determined by the pair of _ complementary _ factor congruences given by the canonical projections @xmath14 .
we prove that factor congruences and direct representations can be described by generalized notions of order ideal and of direct sum of ideals . when the semilattice has a minimum ( maximum ) element , these generalized ideals turn into ordinary ( dual ) ideals .
the main feature of this characterization is that it is completely analogous to similar definitions in the realm of classical algebra , as in ring theory . by using the partial @xmath15 operation
we devise a definition with the same ( quantifier ) complexity as the reader may encounter in classical algebra , and we obtain an equational relation between each element and its direct summands in a given decomposition . by the end of the paper
we apply the characterization to the bounded case and show that each of the axioms is necessary .
the key idea in our characterization stems from the fact that in a direct product of join - semilattices there must exist `` non - trivial '' meets satisfying certain modularity and absorption laws with respect to join . and
conversely , the existence of meets in a direct product implies the existence of them in each factor .
we state without proof the latter property : [ claim : meet_prod ] for every @xmath16-semilattices @xmath17 and elements @xmath18 and @xmath19 , if @xmath20 exists in @xmath21 , then @xmath22 and @xmath23 exist ( and conversely ) . as an a example , a semilattice @xmath6 is pictured in figure [ fig : paraguas ] .
@xmath6 is isomorphic to the direct product of its subsemilattices @xmath24 and @xmath25 .
according to this representation , the pair @xmath26 corresponds to @xmath27 . in spite of @xmath27
not being expressible in terms of @xmath28 , @xmath29 and @xmath5 , we can recover @xmath27 as the infimum @xmath30 .
other relations between @xmath27 , @xmath28 , @xmath29 and @xmath31 ( the only element in @xmath32 ) may be found ; next we choose four of them that hold exactly when @xmath27 corresponds to @xmath26 in a direct decomposition .
isomorphic to @xmath33 . ] in the following , we will write formulas in the extended language @xmath34 . the formula `` @xmath35 '' will be interpreted as `` the infimum of @xmath36 exists and equals @xmath37 '' : @xmath38 unless explicitly indicated ; in general , every equation @xmath39 involving @xmath15 will be read as `` if either side exists , the other one also does and they are equal '' .
it is easy to see that the associative law holds for the partial @xmath15 operation in every @xmath5-semilattice , and we will apply it without any further mention . heretofore , @xmath6 will be a @xmath16-semilattice and @xmath40 arbitrary but fixed .
let @xmath41 be the conjunction of the following formulas : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * @xmath42 * @xmath43 * @xmath44 * @xmath45 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [ defn : direct_sum ] assume @xmath46 are subsemilattices of @xmath6 .
we will say that @xmath6 is the _
@xmath31-direct sum _ of @xmath24 and @xmath25 ( notation : @xmath47 ) if and only if the following hold : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * for all @xmath48 , @xmath49 and @xmath50 , if @xmath51 then @xmath52 * for all @xmath48 , @xmath49 and @xmath50 , if @xmath53 then @xmath54 for @xmath55 .
* for all @xmath56 and @xmath57 , we have : @xmath58 ( and interchanging @xmath24 and @xmath25 ) . * @xmath59 * @xmath60 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in order to make notation lighter
, we will drop the subindex `` @xmath31 '' .
let us notice that * exi * implies : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * @xmath61 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
[ l : suma_dir ] assume @xmath62 .
then @xmath63 defines an isomorphism @xmath64 between @xmath65 and @xmath6 .
first we ll see that the function @xmath66 is well defined .
let us suppose @xmath67 and @xmath68 ( there is at least one possible image by * exi * ) .
we operate as follows : @xmath69 { \wedge}\\ & \quad \quad { \wedge}[z { \vee}((x { \vee}x_2 ) { \wedge}(c { \vee}x_2 ) ) ] & & \text{by { $ \mathbf{p_1}$ } { } and { $ \mathbf{p_2}$}for $ x$.}\end{aligned}\ ] ] by * join * for @xmath27 we may apply * mod2 * : @xmath70 { \wedge}[(z { \vee}x { \vee}x_2 ) { \wedge}(\underline{c { \vee}z } { \vee}x_2)]\ ] ] and by * join * for @xmath37 , @xmath71 { \wedge}[(x { \vee}z { \vee}x_2 ) { \wedge}(x_1 { \vee}x_2)]\ ] ] this last term is symmetric in @xmath27 and @xmath37 , hence we obtain @xmath72 .
the function thus defined by @xmath73 is surjective by * onto * ; let us now prove that it is 11 .
if @xmath74 and @xmath75 , we have : @xmath76 hence @xmath77 .
symmetrically , @xmath78 and in conclusion @xmath79 on the other hand , @xmath80 symmetrically , @xmath81 and hence @xmath82 .
collecting this with ( [ eq : 1 ] ) and using @xmath83 we have @xmath84 by the same reasoning , @xmath85 .
we will now prove that @xmath73 preserves @xmath5 .
let us suppose that @xmath86 and @xmath87 ; since each of @xmath46 is a subsemilattice , we know that @xmath88 for @xmath89 .
we have to see that @xmath90 .
the property * join * is immediate .
we now prove * dist * : @xmath91 by * dist * for @xmath37 and * mod1*. note that @xmath92 hence we can eliminate two terms in equation ( [ eq : 5 ] ) and we obtain * dist * for @xmath93 as expected : @xmath94 we may obtain @xmath83 and @xmath95 similarly .
we prove @xmath83 : @xmath96 by @xmath83 for @xmath37 followed by * mod2 * in each term of the meet , @xmath97 and this equals @xmath98 since the last term in ( [ eq : 6 ] ) is less than or equal to the previous ones .
since @xmath99 defines an isomorphism ( relative to @xmath46 ) , there are canonical projections @xmath100 with @xmath89 such that @xmath101 holds .
[ th : bijection ] let @xmath6 be a @xmath5-semilattice and @xmath102 arbitrary .
the mappings @xmath103 where @xmath104 , are mutually inverse maps defined between the set of pairs of complementary factor congruences of @xmath6 and the set of pairs of subsemilattices @xmath46 of @xmath6 such that @xmath105 .
we first prove that map @xmath106 is well defined , i.e. @xmath107 for every pair of complementary factor congruences @xmath108 .
it s clear that @xmath109 are subsemilattices of @xmath6 , and we know that the mapping @xmath110 is an isomorphism between @xmath6 and @xmath111 . under this isomorphism , @xmath112 corresponds to @xmath113 and @xmath114 to @xmath115 , where @xmath116 and @xmath117 .
henceforth we will identify @xmath112 and @xmath114 with their respective isomorphic images and we will check the axioms for @xmath107 in @xmath118 . in order to see * abs * , let @xmath119 , @xmath120 and @xmath121 as in the hypothesis and suppose that @xmath122 exists .
that is to say , @xmath123 = { \langle}x , c''{\rangle}{\wedge}{\langle}y { \vee}c ' , c '' { \vee}z{\rangle}\ ] ] exists . by claim [ claim : meet_prod ] , we know that @xmath124 must exist in @xmath125 , and in conclusion , @xmath126 = { \langle}x { \wedge}(y { \vee}c ' ) , c '' { \wedge}(c '' { \vee}z){\rangle}= \\ = { \langle}x { \wedge}(y { \vee}c ' ) , c''{\rangle}= { \langle}x , c''{\rangle}{\wedge}[{\langle}y , c''{\rangle}{\vee}{\langle}c ' , c''{\rangle } ] . \end{gathered}\ ] ] that is to say , the other meet @xmath127 exists and equals the former .
the other half is analogous . to check * mod1 * , we assume @xmath128 , @xmath129 , @xmath130 and @xmath131 . note that @xmath51 implies @xmath132 and @xmath133 , and hence we have @xmath134 and @xmath135 . by applying claim [ claim : meet_prod ]
we obtain : @xmath136 the verification of * mod2 * is similar .
let @xmath128 , @xmath137 and @xmath138 .
the map @xmath139 given by @xmath140 is a bijection .
hence , to check * exi * and * onto * we just have to check that @xmath141 holds : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * we have : @xmath142 { \wedge}[{\langle}x ' , x''{\rangle}{\vee}{\langle}c ' , x''{\rangle}]\\ & = ( x { \vee}x_1 ) { \wedge}(x { \vee}x_2 ) , \end{aligned}\ ] ] which is what we were looking for .
* we reason as follows @xmath143 { \wedge}[{\langle}c ' , c''{\rangle}{\vee}{\langle}x ' , c''{\rangle}]\\ & = ( x { \vee}x_1 ) { \wedge}(c { \vee}x_1 ) , \end{aligned}\ ] ] * it s totally analogous . * obvious . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ let @xmath144 and @xmath145 be the projections from @xmath146 onto @xmath112 and @xmath114 , respectively , as defined by ( [ eq:1 ] ) . since @xmath147 , we have @xmath148 for all @xmath149 and @xmath150
. we will prove now that each one of the maps @xmath106 , @xmath151 is the inverse of the other .
we first prove @xmath152 .
for this we have to show that @xmath153 and @xmath154 .
let @xmath155 and @xmath156 . by using ( [ eq:2 ] )
we have : @xmath157 i.e. , @xmath158 .
the proof for @xmath154 is entirely analogous .
now we show @xmath159 .
we check @xmath160 . by taking @xmath161 in * onto
* we can deduce @xmath162 , @xmath163 and hence @xmath164 . by inspection
it is immediate that @xmath165 for all @xmath166 ; in particular , for @xmath167 .
hence we have @xmath168 for all @xmath169 and then @xmath170 . now take @xmath171 ; we have @xmath172 and then @xmath173 for some @xmath174 . by *
join * we have @xmath175 . now considering @xmath83 and * dist
* we obtain : @xmath176 therefore @xmath167 , proving @xmath177 .
we can obtain @xmath178 similarly .
these results can be easily extended for the case of semilattices @xmath179 with _ operators _ , i.e. , operations on @xmath6 preserving @xmath5 : @xmath180 .
we only have to include in the definition of @xmath181 some preservation axioms analogous to * mod1 * and * mod2*. in the case of one unary operator , these are : 1 .
@xmath46 are @xmath182-subalgebras , 2 . for all @xmath183 , @xmath49 and @xmath50 , if @xmath51 then @xmath184 3 . for all @xmath183 , @xmath49 and @xmath50 , if @xmath185 then @xmath186 for @xmath55
if @xmath182 is just a monotone operation , we may also recover the theorem [ th : bijection ] by adding : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for all @xmath183 , @xmath49 and @xmath50 , if @xmath187 then @xmath188 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it is well - known that semilattices have the property of unique factorizability . in general ,
connected posets have the _ refinement property _ : every two finite direct decompositions admit a common refinement .
this was proved by hashimoto @xcite .
furthermore , chen @xcite showed that semilattices have a `` strong form '' of unique factorizability , namely the _ strict refinement property ( srp ) _ ( see , for instance , @xcite ) .
this property is equivalent to the fact that the factor congruences form a boolean algebra . by using our representation of direct decompositions
it can be proved that semilattices have the srp .
we now translate the lattice operations on congruences to the set of `` direct summands '' of @xmath6 , @xmath189 .
set intersection of ideals is the meet . to describe the join , consider two pairs of ideals @xmath190 . then , there exists a subsemilattice @xmath191 of @xmath6 such that @xmath192 .
namely , if we assume ( by using theorem [ th : bijection ] ) that @xmath193 , @xmath194 and @xmath195 , then @xmath196
in the case our semilattice @xmath6 has a 0 or a 1 minimum or maximum element , resp. the characterization of factor congruences by generalized ideals takes a much simpler form . by taking @xmath31 to be the minimum ( maximum ) , it turns out that the subsemilattices @xmath24 and @xmath25 are order ( dual ) ideals .
[ th : caract_0 ] let @xmath6 be a semilattice with 0
. then @xmath197 if and only if @xmath198 satisfy : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1 . for all @xmath56 and @xmath57
, we have : @xmath199 ( and interchanging @xmath24 and @xmath25 ) .
@xmath200 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ moreover , @xmath24 and @xmath25 are ideals of @xmath6 . since @xmath201 , *
abs * is equivalent to its new form as stated ; and * mod1 * , * mod2 * are trivially true .
the * join * part of @xmath202 reduces to `` @xmath203 '' and from here we conclude that * dist * , @xmath83 and @xmath95 hold trivially .
hence * exi * simplifies to @xmath204 which holds trivially , and * onto * reduces to : @xmath205 and this is equivalent to @xmath200 to see that both of @xmath24 and @xmath25 are ideals , we observe that by theorem [ th : bijection ] each of them is the @xmath206-class of @xmath207 for some congruence @xmath206 , hence they are downward closed .
[ th : caract_1 ] let @xmath6 be a semilattice with 1 . then @xmath208 if and only if @xmath198 satisfy : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1 .
@xmath209 2 .
@xmath211 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ moreover , @xmath24 and @xmath25 are dual ideals ( viz . upward closed sets ) of @xmath6 .
we may eliminate * abs * and * mod2 * since they are trivial when @xmath212 ( this can be seen by considering * ori * after definition [ defn : direct_sum ] ) .
formulas @xmath83 and @xmath95 turn into @xmath213 which are equivalent to say @xmath214 . using this ,
we reduce * dist * to : @xmath215 and then we may eliminate @xmath83 and @xmath95 in favor of this last formula . *
join * is equivalent to @xmath216 , and hence * exi * reduces to the statement `` for all @xmath217 , @xmath216 and @xmath218 exists '' .
we may state the first part as @xmath219 , and may condense the second part with * onto * writing @xmath210 .
it remains to take care of * mod1*. the hypothesis @xmath51 trivializes when @xmath212 .
it is obvious that * mod1 * implies * mod1 * , by taking @xmath220 .
and we may obtain the new version under the hypotheses of the theorem by taking @xmath221 in * mod1*.
we check independence and necessity by providing semilattices with a distinguished element @xmath31 and a pair of ideals @xmath24 and @xmath25 such that they satisfy every axiom except one and it is not the case that @xmath222 .
it is immediate that the minimal non - modular lattice @xmath223 provide two such counterexamples showing that * abs * and * mod1 * are necessary and independent ( see figure [ fig : ejemplos ] ( a ) and ( b ) , respectively ) . in figure
[ fig : ejemplos ] ( a ) we have taken @xmath31 to be the minimum element , hence theorem [ th : caract_0 ] ( and its proof ) applies and we only should check * onto * , which is obvious .
nevertheless , * abs * fails for the labeled elements @xmath28 , @xmath224 and @xmath225 . note that * abs * and * onto * imply @xmath226 , but this sole assumption does not suffices to prove uniqueness of representations : in this example we have @xmath227 and @xmath228 . in figure
[ fig : ejemplos ] ( b ) @xmath31 is the maximum of the semilattice , hence by theorem [ th : caract_1 ] ( and its proof ) we only have to check * exi * and * onto * , again obvious , and * mod1 * fails for this model . | we characterize factor congruences in semilattices by using generalized notions of order ideal and of direct sum of ideals . when the semilattice has a minimum ( maximum ) element , these generalized ideals turn into ordinary ( dual ) ideals . en
este trabajo damos una caracterizacin de las congruencias factor en semiretculos usando nociones generalizadas de ideal y suma directa de ideales .
cuando un semiretculo tiene elemento mnimo ( mximo ) , estos ideales generalizados resultan ideales ( duales ) ordinarios . |
laser radiation in the form of strong and short pulses finds nowadays application in many branches of physical research .
the most popular example is the interaction of strong and short pulses with atomic and molecular systems , and in particular ionization of atoms and molecules . for strong fields with maximum intensities not exceeding one atomic unit ( @xmath4 )
dipole approximation is sufficient to account for major features of atomic and molecular ionization @xcite . in such cases
kinetic energy of the ejected electron is small compared to its rest energy , so that motion of the photoelectron can be described in the framework of non - relativistic dynamics in a plane - wave electromagnetic field @xcite .
however , the dipole and plane wave approximation fails in the case of ultrastrong tightly focused pulses , when spatial dependence of the field becomes important @xcite and non - relativistic description of the electron motion becomes inadequate . in the purely classical approach description of focused laser radiation can be based on various forms of three dimensional non - paraxial approximation @xcite .
solutions of maxwell equations of this type were recently applied to the description of xe - photoionization in ultrastrong fields @xcite characterized by the values of parameters for which the standard one - dimensional dipole approximation is not justified . with the advent of ultrastrong short pulses of radiation for which the dipole approximation does not provide a sufficient theoretical description , it is desirable to formulate a theoretical approach accounting both for the propagation effects and the transverse structure of the pulse .
one of the possibilities is given by non - paraxial description mentioned above , which is based on approximate solution of maxwell equations in classical electrodynamics without taking the quantum effects in electromagnetic field into account .
the purpose of this paper is to propose fully quantum description of spatial and temporal dependence of an ultrastrong pulse of radiation by means of coherent states .
although to a good approximation laser radiation can be described by classical solutions of maxwell equations , its quantum nature is also important . in particular , the use of coherent states allows not only to analyze space images of the pulse field ( as in the purely classical approach ) , but also to determine directly distribution of photon modes in the pulse .
this is not possible in the case of a purely classical soluion , where the photons are simply `` absent '' , and , among various states of quantum electromagnetic field the coherent state gives a good approximation to the real laser field @xcite .
coherent states have been also used for the description of light beams in a recent publication @xcite dealing with the uncertainty relation for light beams .
electromagnetic field of a pulse of radiation is localized in a bounded region of space moving with the speed of light in the direction of pulse propagation .
exact classical solutions of maxwell equations describing focused or otherwise bounded configurations of electromagnetic field have been discussed in the past @xcite . on the other hand , it has been shown by i. biaynicki - birula and z. biaynicka - birula that , in general , the photons `` can not be sharply localized '' @xcite , with the only exception given by coherent states , when full localization of photons is possible . in general , quantum limitations on the localization of energy density do not apply to quantum states having classical counterparts @xcite , and coherent state is a well known example . for these reasons coherent states of quantum electromagnetic field offer , as will be seen below , an adequate description of pulsed laser radiation .
the main idea of the present paper is to construct the coherent state using space - dependent annihilation and creation operators , introduced for the first time by glauber @xcite and applied recently @xcite in the discussion of photon localization problem .
the advantage of using space - dependent operators instead of the standard monochromatic mode operators is that the approach of this type allows to obtain in a simple and direct way space and time profiles of the electric and magnetic fields of the pulse beyond dipole approximation . on the other hand
, an important question is in what way one can determine the distribution of photons inside the pulse .
the answer to this question requires transition from the space - time description to the , complementary to it , photon mode description , or , in other words , transition from the field to the photon picture of the pulse . it can be also considered as the question of transition from the wave - packet modes to the monochromatic modes of quantum electromagnetic field , discussed in the paper by smith and raymer @xcite .
this approach can be also used to find the photon - mode structure of the pulse of radiation .
the main characteristics of the coherent state describing the pulse of radiation is its profile @xmath5 in the configuration space @xcite and its fourier transform in the wave - vector space @xmath6 .
the latter will be assumed in the gaussian form both in the longitudinal direction ( direction of propagation ) and in the transverse direction .
the width of the profile in the wave - vector space in the longitudinal direction will be assumed larger than in the transverse plane of the pulse , which corresponds to a pulse with transverse dimension in the configuration space larger than in the longitudinal direction .
these assumptions about the gaussian profile of the pulse allowed for the approximate analytic calculation of the expectation values of electric and magnetic field in the coherent state describing pulsed laser radiation .
expectation values of the fields were further analyzed as functions of coordinates and time , and represented graphically in the form of contour plots in the @xmath7 plane , where @xmath0 is the direction of propagation of the pulse and @xmath1 is the direction of linear polarization . in general ,
the electric field has two components ; the dominant @xmath1-component and the longitudinal component in the direction of propagation , so that the pulse does not constitute a purely transverse electromagnetic wave , although the longitudinal part is a few orders of magnitude smaller than the transverse one . the dominant component of the magnetic field is in the @xmath3-direction and is accompanied by small components in the @xmath0- and @xmath1-directions .
the non - transverse contributions to electromagnetic field of the pulse vanish after sufficiently long time , when the pulse approaches a linearly polarized plane wave .
the paper is organized as follows . in section 2 , which has a review character ,
i give a brief summary of the formalism of time and coordinate dependent annihilation and creation operators based on @xcite .
section 3 contains expressions for the expectation values of the vector potential and discussion of the photon modes of the coherent state describing the pulse .
the space - time structure of the pulse is discussed in details in section 4 , which also contains graphical representations of the space images of the electric and magnetic field .
section 5 contains final remarks and some details of the calculations have been moved to appendices .
in this review part i briefly summarize for the sake of completeness the formalism of space dependent annihilation and creation operators , presented in ref @xcite .
the operators of quantum electromagnetic field have the following decomposition into the annihilation and creation parts , analogous to the positive and negative frequency decomposition [ eq:1 ] @xmath8,\ ] ] @xmath9,\ ] ] where the electric annihilation and creation operators are denoted by @xmath10 and @xmath11 respectively , and the corresponding magnetic operators by @xmath12 and @xmath13 .
the electric and magnetic operators are not independent and fulfill the relations @xmath14 where @xmath15 is the helicity operator which in the wave - vector space acts as @xmath16 , and in the position space its action on a vector function @xmath17 looks as @xcite @xmath18 the equal time commutation relations between components of the @xmath19 , and @xmath20 operators have a nonlocal character [ eq:4 ] @xmath21=\left(\partial_i\partial_j-\delta_{ij}\delta\right)\frac{1}{2\pi^2|\mathbf{r}-\mathbf{r}'|^2},\ ] ] @xmath22=\left(\partial_i\partial_j-\delta_{ij}\delta\right)\frac{1}{2\pi^2|\mathbf{r}-\mathbf{r}'|^2},\ ] ] whereas the commutator of @xmath23 and @xmath24 is local @xmath25=-i\epsilon_{ikj}\partial_k\delta^{(3)}(\mathbf{r}-\mathbf{r}').\ ] ] the remaining commutators vanish . + due to nonlocal character of the commutation relations the fock basis costructed with the use of either the electric or magnetic creation operators are not orthogonal , and both basis are mutually orthogonal due to the local commutator @xcite .
coherent state is defined with the use of space - dependent annihilation and creation operators as @xmath26|0\rangle,\ ] ] where @xmath27 is the vacuum state and the space profile of the coherent state @xmath28 is given by a divergenceless complex vector function .
coherent state is the eigenstate of annihilation operator @xmath29 where @xmath30 is the fourier transform of the space profile of the coherent state , @xmath31 eq .
can be easily obtained from the definition of coherent state , using the commutation relation . in the similar way @xmath32
this section is devoted to the description of electric and magnetic field of the coherent pulse , and also to the discussion of its photon modes .
electromagnetic field of the pulse is given by expectation values , [ eq:10 ] @xmath33 @xmath34 to find the expectation values at an arbitrary time it is necessary to know action of the annihilation operator @xmath35 on the coherent state . using the heisenberg equations of motion , @xmath36\ ] ] with the hamiltonian @xmath37,\ ] ] one obtains with the use of eqs .
@xmath38 where @xmath39 denotes scalar product of the four - vectors @xmath40 and @xmath1 in the minkowski space , and @xmath41 substitution of into gives @xmath42 and the corresponding formula for the magnetic annihilation operator reads @xmath43 using the decomposition of the field operators into space dependent annihilation and creation operators and the formulae and it is straightworward to calculate the expectation values , @xmath44 where the vector potential has the form @xmath45.\ ] ] obviously , a result of this type should be expected , since , in principle , any solution of the wave equation can be written in this form . however , by using the decomposition and the relations and one can retrieve the correct overall factor and also determine correctly distribution of photons in the pulse .
solution of classical wave equation without quantum attributes of the coherent state does not give such possibilities . at most
, the photon mode distribution can be in this case determined only in a heuristic way . to determine the photon mode distribution of the pulse it is necessary to `` translate '' the space - time description of the coherent state into the photon mode description .
the standard annihilation and creation operators of the monochromatic photon modes are normalized to fulfill the commutation relation @xmath46 = 2(2\pi)^3|\mathbf{k}|\delta^{(3)}(\mathbf{k}-\mathbf{k}')\delta_{\lambda\lambda ' } , \ ] ] where @xmath47 is the polarization index . with this normalization
the decomposition of the electric field operator into positive and negative frequency parts has the form @xcite @xmath48 where @xmath49^{-1}d^3k$ ] is the invariant measure on the light cone in the reciprocal space and @xmath50 is the polarization vector .
annihilation operator of the wave packet mode is given by @xcite @xmath51 where the wave packet profiles fulfill the relations of completness and orthogonality [ eq:22 ] @xmath52 @xmath53 which implies the usual commutation relation for the annihillation and creation operators of the wave packet modes @xmath54=\delta_{\alpha\beta}.\ ] ] the formula inverse to has the form @xmath55 the coherent state ( [ eq:6 ] ) can be written with the use of the wave packet annihillation and creation operators ( [ eq:21 ] ) as @xmath56 where the quantities @xmath57 are expressed by the coherent state profiles @xmath58 and the wave packet profiles @xmath59 as @xmath60 details of the derivation of eqs .
( [ eq:25 ] ) and ( [ eq:26 ] ) are given in the appendix [ a ] .
the photon number operator is given by @xmath61 and its average value in the coherent state @xmath62 , @xmath63 , can be also expressed explicitly by the profile of the coherent state @xmath64 it follows immediately from ( [ eq:27 ] ) that the distribution of polarized photons in the wave vector space is given by @xmath65 and in the case of unpolarized photons @xmath66 expression ( [ eq:29 ] ) can be also obtained by calculating energy distribution in the wave vector space using the hamiltonian ( [ eq:12 ] ) .
its average in the coherent state is given by @xmath67 so that the energy distribution reads @xmath68 and the photon number distribution , @xmath69 agrees with ( [ eq:29 ] ) .
one could also construct the coherent state in standard way using directly the annihilation and creation operators of the photon modes , @xmath70 , where @xmath71 the profiles @xmath58 and @xmath72 in wave vector space are related by @xmath73,\ ] ] which implies @xmath74 .
to discuss the space structure of the pulse in more details specific form of the coherent state profile has to be chosen . the profile
@xmath75 is a divergenceless vector so that @xmath76 .
a simple form of the coherent state profile fulfilling the transversality condition is given by @xmath77 where @xmath78 is a constant unit vector , which in general may be complex , and @xmath79 is a function of the wave vector .
the vector @xmath80 is an analog of fourier transform of positive frequency part of the `` superpotential '' @xmath81 introduced in ref .
@xcite . with the use of ( [ eq:32 ] )
the vector potential ( [ eq:18 ] ) can be written as @xmath82,\ ] ] where @xmath83 according to ( [ eq:27 ] ) the normalization condition for the profile of coherent state is @xmath84 where @xmath85 is the shorthand notation for the mean number of photons @xmath86 .
the number of photons in the pulse determines its intensity and the intensity of its electric field as well . in terms of the function
@xmath87 the normalization condition reads @xmath88 the function @xmath79 does not need to be regular for @xmath89 .
the only requirement is that that the integral determining @xmath90 be convergent . obviously , also the integral in ( [ eq:36 ] ) must exist , but due to higher positive powers of @xmath91 in the integrand formula ( [ eq:34 ] ) imposes stronger limitations on the behavior for small wave vectors .
the scalar profile @xmath87 will be assumed in the factorized form @xmath92 with the function @xmath93 , called also the longitudinal profile , strongly peaked at @xmath94 and the transverse profile @xmath95 significantly narrower than @xmath96 .
the profile of this type corresponds in the configuration space to a pulse moving in the @xmath0-direction with the electric field oscillating with the central frequency @xmath97 , and with the longitudinal size significantly smaller than the transverse one .
the factor @xmath98 will be assumed in the form @xmath99 , where the power @xmath100 can be negative .
it is shown in appendix b that for the integral in ( [ eq:34 ] ) to exist this parameter must fulfill the inequality @xmath101 . for practical reasons only integer powers will be kept , either @xmath102 or @xmath103 . whereas for a strongly peaked longitudinal profile the choice of @xmath100 does not have a significant impact on the shape of @xmath90 , it is important for the shape of the photon number distribution ( [ eq:29 ] ) , especially at the long wavelength edge .
@xmath104 is the normalization constant to be determined from ( [ eq:36 ] )
. the transverse profile will be assumed in the gaussian form @xmath105 where @xmath106 is a parameter determining width of the transverse profile ; its inverse is of the order of the transverse width of the pulse in the configuration space .
form of the longitudinal profile @xmath96 depends on the shape of the pulse envelope . for a gaussian envelope it is also given by a gaussian . in the sfa - type calculations the commonly used model of the pulse
has the cosine squared - envelope @xcite which obviously corresponds to a non - gaussian profile in the wave vector space .
if width of the longitudinal profile @xmath93 ( e.g. full width at half - maximum ) is of the order of @xmath107 electromagnetic field of the pulse is confined in a region of the volume @xmath108 moving with the velocity of light . for a linearly polarized pulse
the vector @xmath78 can be chosen as @xmath109 , and the normalization condition ( [ eq:36 ] ) takes then the form @xmath110 where @xmath111 is the azimuthal angle in cylidrical coordinates in the @xmath112-space .
writing the normalization factor in the form @xmath113 and performing angular integration reduces ( [ eq:39 ] ) to @xmath114 where @xmath115 will be taken either as @xmath102 or @xmath116 .
detailed calculation of the normalization constant is given in appendix b. using the formula ( [ eq:37 ] ) for @xmath79 with transverse profile of the form ( [ eq:38 ] ) , and the longitudinal profile strongly peaked for @xmath94 , one can write the approximate expression for the function @xmath90 @xmath117 where [ eq:42 ] @xmath118 @xmath119 and @xmath120 .
details of this approximate calculation are given in appendix [ b ] , where it is also shown that the factor @xmath121 is finite and non - zero in the plane wave limit @xmath122 . in general , the integration in ( [ eq:34 ] ) and ( [ eq:39 ] ) could be done analytically under the assumption that dominant contribution to the integral comes from the region in the @xmath112-space for which @xmath123 . in the standard approach to the theoretical description of interaction with strong pulsed fields
the pulse is described by a plane wave with the functional dependence on @xmath0 and @xmath124 of the type @xmath125 ( z - dependence is suppressed in the dipole approximation ) .
this formally corresponds to the limit @xmath126 , such that @xmath127 .
this last condition allows to keep finite the number of photons per unit volume in which the electromagnetic field of the pulse is confined .
this volume is of the order of @xmath128 , where @xmath107 is the parameter measuring width of the longitudinal profile @xmath96 .
it can be then seen from ( [ eq:41 ] ) that the expression for @xmath90 has a well definite plane wave limit .
however , it has to be noted that , strictly speaking , the notion of local photon density in space ( described by a function of coordinate @xmath129 ) can not be used @xcite .
it follows from ( [ eq:33 ] ) that @xmath90 must be a dimensionless quantity and , with the choice of the longitudinal profile @xmath96 also as a dimensionless function , it follows from ( [ eq:41 ] ) that the dimension of the factor @xmath121 is @xmath130 , which agrees with ( [ eq:40 ] ) .
since @xmath96 is strongly peaked at @xmath94 it is convenient to shift the integration variable in ( [ eq:42b ] ) , @xmath131 and write the expression for @xmath132 as @xmath133 where @xmath134 for dimensional reasons @xmath135 can be written as @xmath136 with dimensionless @xmath137 , which can be also normalized to unity for @xmath138 , i.e. @xmath139 .
e.g. for a pulse with the gaussian envelope one has @xmath140 and @xmath141 expression ( [ eq:41 ] ) can now be written as @xmath142 where @xmath143 . according to ( [ eq:33 ] )
the vector potential is given by @xmath144 [ cols="^,^ " , ] where @xmath145 .
it follows from ( [ eq:47 ] ) that @xmath146,\ ] ] where @xmath147 @xmath148 is the phase of the function @xmath149 ( [ eq:42a ] ) , and @xmath150 , @xmath151 denote , respectively , the real and imaginary part of @xmath149 .
components of the vector potential read [ eq:51 ] @xmath152+\frac{h'_1(z - ct)}{h_1(z - ct)}\cos[g(\mathbf{r},t+\gamma(t)]\right\},\ ] ] @xmath153.\ ] ] the vector potential and the electric field have non - vanishing longitudinal component , @xmath154 and @xmath155 , which for short pulses with relatively large transverse spatial extension is small as compared to @xmath156 @xcite . in the plane wave limit
the longitudinal component vanishes and the transverse component of the vector potential has the limiting form @xmath157+\frac{h_1'(z - ct)}{h_1(z - ct)}\cos[k_0(z - ct)]\right\},\ ] ] where @xmath158 is a constant quantity in the plane wave limit . defining @xmath159,\ ] ]
one can write ( [ eq:52 ] ) as @xmath160.\end{aligned}\ ] ] for a gaussian pulse @xmath161 . for a typical laser pulse @xmath162 , so that the additional phase in ( [ eq:54 ] ) can be approximated as @xmath163 . approximating also @xmath164 by unity we get @xmath165,\end{aligned}\ ] ] where the shifted wave number of the carrier wave is @xmath166 , and shifted frequency @xmath167 . in the dipole approximation vector potential of the pulse
is usually written as @xmath168 where @xmath169 denotes maximum value of the electric field . by comparison of
( [ eq:55 ] ) and ( [ eq:56 ] ) we can estimate the electric field amplitude in the present case as @xmath170 with the approximation @xmath171 .
the vector potential ( [ eq:51 ] ) can be therefore written as [ eq:58 ] @xmath172 , \end{aligned}\ ] ] @xmath173 , \end{aligned}\ ] ] where [ c.f .
eq.([eq:50 ] ) ] @xmath174 electric field is calculated as @xmath175 .
[ fig1 ] shows space images of the @xmath1- and @xmath0 - components of the electric field of the pulse in the @xmath2 plane at the focus , i.e. for @xmath176 .
the longitudinal component of the field is much smaller than the transverse one and vanishes along the @xmath1-axis .
these properties are similar to those of the pulse modelled by non - paraxial field @xcite .
time dependence of the pulse is described by the functions @xmath177 , @xmath178 , @xmath179 and @xmath148 in ( [ eq:58 ] ) . in general , transverse dimensions of the pulse increase during time evolution .
this can be easily seen from the factor @xmath180 $ ] in ( [ eq:58 ] ) , where the function @xmath178 decreases at large @xmath124 as @xmath181 , and therefore width of the gaussian function grows . with increasing transverse dimensions the density of energy carried by the pulse decreases , and
this is accounted for by the overall factor @xmath177 in ( [ eq:58 ] ) , which decreases as @xmath182 .
one can also note that with increasing time edges of the pulse in the transverse direction lag behind the center of the pulse , see fig .
this effect occurs due character of the time dependence of the function @xmath183 ( [ eq:59 ] ) , and , in particular , to the time dependence of imaginary part of @xmath149 , ie .
@xmath179 , which for times fulfilling @xmath184 is close to its maximum value . for the gaussian pulse
( [ eq:46 ] ) the square bracket in the expression ( [ eq:58a ] ) for the @xmath1-component of the vector potential can be written as @xmath185,\ ] ] where @xmath186 the dimensionless variables and parameters are @xmath187 . for @xmath188 , as in fig .
[ fig1 ] , the phase @xmath111 does not depend on transverse distance @xmath189
. therefore the line of constant phase , where the sine function in ( [ eq:60 ] ) has constant value , is in the @xmath2 plane a straight line parallel to the @xmath1-axis .
for @xmath190 such that @xmath191 the @xmath192 dependence of the phase becomes significant and the values of @xmath193 for which @xmath194 are for @xmath195 smaller than for @xmath196 , which leads to``bending '' of the pulse ; this effect is shown in fig .
[ fig2 ] . for still larger times ,
such that @xmath197 , the phase @xmath111 becomes weakly dependent on the transverse distance @xmath189 .
the pulse approaches then the linearly polarized plane wave with large transverse dimension .
maximum value of the electric field becomes smaller and smaller , since energy carried by the pulse is then spread over larger region .
expectation values of the magnetic field components are given by ( [ eq:17 ] ) , or , explicitly @xmath198 it can be seen from ( [ eq:58 ] ) that the derivatives @xmath199 in ( [ eq:62 ] ) are small compared to @xmath200 .
the @xmath1-component of the vector potential does not contain the small overall factor @xmath201 , as well as its derivative with respect to @xmath0 . in the case of three other derivatives
either @xmath154 contains the @xmath201-factor , or it appears as a result of differentiation in @xmath202 . for this reason
the dominant component of the magnetic field is @xmath203 with other components smaller by few orders of magnitude . in the plane wave limit ( @xmath204 with @xmath127 )
only transverse components of the fields survive .
it has been shown explicitly in this paper how a pulse of laser radiation can be described with the use of the coherent state .
the most convenient theoretical tool used for this decription are the space dependent annihillation and creation operators @xcite instead of the standard operators pertaining to the photon mode approach .
the use of space dependent operators allowed for a direct description of space and time dependence of the electromagnetic field of the pulse ; this was achieved by calculating expectation values of the corresponding field operators .
another important characteristic of the pulse of radiation is distribution of photon modes , which was found using the connection between the field description and photon mode ( wave packet ) description of the quantum electromagnetic field @xcite .
although in principle the space - time dependence of the pulsed field could be determined by constructing an appropriate solution of classical maxwell equation ( e.g by choosing an appropriate wave packet profile in the plane wave decomposition ) , quantum attributes of the coherent state are essential for exact determination of the photon mode distribution .
the use of coherent state allows also to find the exact relation between maximum value of the electric field intensity and the total number of photons carried by the pulse [ c.f .
formula ( [ eq:57 ] ) ] .
although standard description of interaction with strong laser fields relies on the plane wave and dipole approximation , details of the pulse structure , also in the plane transverse with respect to the direction of propagation , become important for very strong pulses with shorter wavelengths of the carrier wave @xcite .
in order to derive eqs .
( [ eq:25 ] ) and ( [ eq:26 ] ) we note that according to ( [ eq:1a ] ) and ( [ eq:20 ] ) space and time dependent annihilation operators @xmath35 can be expressed by annihilation operator of the photon mode as @xmath205 substituting ( [ eq:24 ] ) into ( [ eq : a1 ] ) one obtains @xmath206 defining now @xmath57 according to([eq:26 ] ) we obtain @xmath207 and ( [ eq:25 ] ) follows from ( [ eq:6 ] ) and ( [ eq : a3 ] ) . the normalization condition ( [ eq:40 ] ) can be written in the form @xmath208 where @xmath209 introducing dimensionless variable of integration , @xmath210 one gets @xmath211 the integration can be done explicitly with the use of @xcite @xmath212 where @xmath213 is the confluent hypergrometric function of the second type .
the parameter @xmath214 in ( [ eq : b4 ] ) corresponds to @xmath215 . using asymptotic form of the @xmath213-function for small argument , @xmath216
, one can easily show that the function @xmath217 is finite for @xmath218 if @xmath219 .
this condition also leads to the photon number distribution ( [ eq:29 ] ) regular for @xmath220
. it will be shown below that the convergence requirement for the integral determining the @xmath90 ( [ eq:34 ] ) also imposes the same limitation on @xmath221 . since dominant values of @xmath222 in ( [ eq : b1 ] )
are close to @xmath223 , and @xmath224 , one can use asymptotic form of @xmath213 for large argument , @xmath225 .
further , @xmath217 is a slowly varying function compared to @xmath96 so that , practically , @xmath217 in ( [ eq : b1 ] ) can be replaced by @xmath226 , which for @xmath224 reads @xmath227 and ( [ eq : b1 ] ) can be written as @xmath228 for the gaussian longitudinal profile ( [ eq:45 ] ) one gets @xmath229 substituting ( [ eq:37 ] ) into ( [ eq:34 ] ) and performing angular integration in cylindrical coordinates one gets @xmath230 where @xmath231 and @xmath232 is the bessel function of zero order .
the function @xmath233 is regular at @xmath234 for @xmath235 , and , assuming nonnegative and integer values of @xmath221 , the possibilities are @xmath236 and @xmath237 .
since the longitudinal profile @xmath96 is strongly peaked for @xmath94 and the integral over @xmath238 in ( [ eq : b9 ] ) is dominated by small values of @xmath238 of the order of @xmath106 , the main contribution to the integral comes from the region of @xmath40-space in which @xmath239 .
one can therefore replace the slowly varying factor @xmath240 by @xmath241 , and use an expansion in the oscillating exponential function in ( [ eq : b9 ] ) @xmath242 so that @xmath243j_0(k_\perp r_\perp),\end{aligned}\ ] ] where @xmath244 the integration can be performed explicitly using the formula @xmath245 which gives @xmath246.\end{aligned}\ ] ] substituting ( [ eq : b14 ] ) into ( [ eq : b8 ] ) we obtain @xmath247.\end{aligned}\ ] ] integration over positive @xmath222-axis , when @xmath248 , gives the wave propagating in the positive @xmath0 direction and the integral over negative values of @xmath222 gives rise to a wave propagating in the negative @xmath0 direction .
since the profile @xmath96 is strongly peaked at @xmath249 , and rapidly decreases far off @xmath94 , the counter - propagating wave has very small amplitude and can be disregarded .
one can therefore put @xmath248 and extend the integration over entire range of @xmath222 from @xmath250 to @xmath251 , since the contribution from negative values of @xmath222 is very small whatsoever .
disregarding further the @xmath222-dependence of the slowly varying function @xmath252 , ie . replacing @xmath253 in the integrand gives expression ( [ eq:41 ] ) and further ( [ eq:47 ] ) for the function @xmath90 .
miloevi , g.g .
paulus , d. bauer , w. becker , j. phys .
b : at . mol .
. phys . * 39 * r203 ( 2006 ) .
dichiara , i. ghebregziabher , j.m .
waesche , t. stanev , n. ekanayake , l.r .
barclay , s.j .
wells , a. watts , m. videtto , c.a .
manusco , b.c .
walker , phys .
a * 81 * 043417 ( 2010 ) .
m. lax , w.h .
louisell , w.b .
mcknight , phys .
rev . a * 11 * 1365 ( 1975 )
davis , phys .
rev . a * 19 * 1177 ( 1979 ) .
huayu hu and jianmin yuan , phys . rev .
a * 78 * , 063826 ( 2008 ) .
i. biaynicki - birula , z. biaynicka - birula , phys .
a * 88 * 052103 ( 2013 ) r. w. hellwarth , p. nouchi , phys .
e * 54 * 889 ( 1996 ) .
r. w. ziolkowski , phys .
a * 39 * 2005 ( 1989 )
. i. biaynicki - birula , phys .
letters * 80 * 5247 ( 1998 ) .
i. biaynicki - birula , z. biaynicka - birula , j. phys .
b : at . mol .
. phys . * 39 * s545 ( 2006 ) .
i. biaynicki - birula , z. biaynicka - birula , phys .
a * 79 * 032112 ( 2009 ) .
r. j. glauber , phys .
rev . * 130 * 2529 ( 1963 ) .
r. j. glauber , phys .
rev . * 131 * 2766 ( 1963 ) .
b. j. smith , m. g. raymer , new j. phys .
* 9 * 414 ( 2007 ) .
i. biaynicki - birula , z. biaynicka - birula , _ quantum electrodynamics _ ( pergamon , oxford , 1975 ) , p. 163 .
i. biaynicki - birula , z. biaynicka - birula , phys .
a * 86 * 022118 ( 2012 ) .
nist digital library of mathematical functions , http://dlmf.nist.gov/13.4.4 | space - time description of pulsed laser radiation by means of coherent states is presented .
the corresponding displacement operator contains space dependent annihilation and creation operators , not the `` standard '' operators corresponding to mode decomposition of the electromagnetic field .
this allowed for a direct description of space and time dependence of the electromagnetic field of the pulse .
the main characteristics of the pulse is its profile in the wave - vector space .
it determines space - time structure of expectation values of the electromagnetic field carried by the pulse , and the distribution of photon modes as well .
determination of the latter required transition from the space - time , or field description to the photon mode approach .
expectation values of the electric field of a gaussian pulse propagating in the @xmath0-direction and linearly polarized in the @xmath1-direction are presented graphically in the form of contour plots in the @xmath2 plane .
apart of the dominant @xmath1-component electric field has a small longitudinal component in the @xmath0-direction .
the dominant component of the magnetic field in the @xmath3-direction is acompannied by small components in @xmath0- and @xmath1-directions .
the plane wave limit of the pulse can be implemented by shrinking transverse width of the profile to zero ( transverse spatial extension goes to infinity ) , while keeping the photon density constant .
one can then retrieve standard plane wave form of the pulse also in the dipole approximation . |
analyzing whether a time series is stationary or is a non - stationary random walk ( unit root process ) in the sense that the first order differences form a stationary series is an important issue in time series analysis , particularly in econometrics .
often the task is to test the unit root null hypothesis against the alternative of stationarity at a pre - specified @xmath0 level , which ensures that a decision in favor of stationarity is statistically significant .
for instance , the equilibrium analysis of macroeconomic variables as established by granger ( 1981 ) and engle and granger ( 1987 ) defines an equilibrium of two random walks as the existence of stationary linear combination .
when analyzing equilibrium errors of a cointegration relationship , rejection of the null hypothesis in favor of stationarity means that the decision to believe in a valid equilibrium is statistically justified at the pre - specified @xmath1 level . for an approach where cusum based residual tests are employed to test the null hypothesis of cointegration , we refer to xiao and phillips ( 2002 ) .
their test uses residuals calculated from the full sample . in the present article
we study sequential monitoring procedures which aim at monitoring a time series until a time horizon @xmath2 to detect stationarity as soon as possible .
the question whether a time series is stationary or a random walk is also of considerable importance to choose a valid method when analyzing the series to detect trends .
such procedures usually assume stationarity , see steland ( 2004 , 2005a ) , pawlak et al .
( 2004 ) , huskov ( 1999 ) , huskov and slab ( 2001 ) , ferger ( 1993 , 1995 ) , among others . as shown in steland ( 2005b ) ,
when using nadaraya - watson type smoothers to detect drifts the limiting distributions for the random walk case differ substantially from the case of a stationary time series . to detect changes in a process or
a misspecified model , a common approach originating in statistical quality control is to formulate an in - control model ( null hypothesis ) and an out - of - control model ( alternative ) , and to apply appropriate control charts resp .
stopping times . given a time series @xmath3 a monitoring procedure with time horizon ( maximum sample size ) @xmath2
is given by a stopping time @xmath4 using the convention @xmath5 , where @xmath6 , called _ control statistic _ , is a @xmath7-measurable @xmath8-valued statistic sensitive for the alternatives of interest , and @xmath9 is a measurable set such that @xmath10 has small probability under the null model and high probability under the alternative of interest . in most cases
@xmath11 is of the form @xmath12 or @xmath13 for some given _ control limit _ ( critical value ) @xmath14 . to design monitoring procedures ,
the standard approach is to choose the control limit to ensure that the average run length ( arl ) , @xmath15 , is greater or equal to some pre - specified value .
however , controlling the significance level is a also serious concern .
the results presented in this article can be used to control any characteristic of interest , although we will focus on the type i error in the sequel . the ( weighted ) dickey - fuller control chart studied in this article
is essentially based on a sequential version of the well - known dickey - fuller ( df ) unit root test , which is motivated by least squares . due to its power properties
this test is very popular , although it is known that its statistical properties strongly depend on a correct specification of the correlation structure of the innovation sequence .
the df test and its asymptotic properties , particularly its non - standard limit distribution have been studied by white ( 1958 ) , fuller ( 1976 ) , rao ( 1978 , 1980 ) , dickey and fuller ( 1979 ) , and evans and savin ( 1981 ) , chan and wei ( 1987 , 1988 ) , phillips ( 1987 ) , among others .
we will generalize some of these results . to ensure quicker detection in case of a change to stationarity
, we modify the df statistic by introducing kernel weights to attach small weights to summands corresponding to past observations .
we provide the asymptotic theory for the related dickey - fuller ( df type ) processes and stopping times , also covering local - to - unity alternatives
. for correlated error terms the asymptotic distribution of the df test statistic , and hence the control limit of a monitoring procedure , depends on a nuisance parameter , which can be estimated by newey - west type estimators .
we consider two approaches to deal with that problem .
firstly , based on a consistent estimate of the nuisance parameter one may take the asymptotic control limit corresponding to the estimated value .
secondly , following phillips ( 1987 ) one may consider appropriate transformations of the processes possessing limit distributions which no longer dependent on the nuisance parameter .
a nonparametric approach called kpss test which avoids this problem , at least for i(1 ) processes , has been proposed by kwiatkowski et al .
that unit root test has better type i error accuracy , but tends to be less powerful .
monitoring procedures related to this approach and their merits have been studied in detail in steland ( 2006 ) .
the organisation of the paper is as follows . in section [ secmodel ]
we explain and motivate carefully our assumptions on the time series model , and present the class of dickey - fuller type processes and related stopping times .
the asymptotic distribution theory under the null hypothesis of a random walk is provided in section [ ash0 ] .
section [ ash1 ] studies local - to - unity asymptotics , where the asymptotic distribution is driven by an ornstein - uhlenbeck process instead of the brownian motion appearing in the unit root case . finally , in section [ secsim ] we compare the methods by simulations .
our results work under quite general nonparametric assumptions allowing for dependencies and conditional heteroskedasticity ( garch effects ) , thus providing a nonparametric view on the parametrically motivated approach . to motivate our assumptions , let us consider the following common time series model , which is often used in applications .
suppose at this end that @xmath16 is an ar(@xmath17 ) time series , i.e. , @xmath18 for starting values @xmath19 , where @xmath20 are i.i.d .
error terms ( innovations ) with @xmath21 and @xmath22 , @xmath23 . assume the characteristic polynomial @xmath24 has a unit root , i.e. , @xmath25 , of multiplicity @xmath26 , and all other roots are outside the unit circle , i.e. , @xmath27 implies @xmath28 .
then @xmath29 for some polynomial @xmath30 with has no roots in the unit circle implying that @xmath31 exists for all @xmath32 .
we obtain @xmath33 , where @xmath34 denotes the lag operator .
since @xmath35 can be inverted , we have the representation @xmath36 for coefficients @xmath37 this means , @xmath16 satisfies an ar(@xmath26 ) model with correlated errors . for the calculation of @xmath38
we refer to brockwell and davis ( 1991 , sec . 3.3 . ) in particular , to analyze an ar(@xmath17 ) series for a unit root , one can work with an ar(@xmath26 ) model with correlated errors .
the representation ( [ arpasar1 ] ) motivates the following time series framework which will be assumed in the sequel .
suppose we are given an univariate time series @xmath39 satisfying @xmath40 where @xmath41 $ ] is a fixed but unknown parameter . concerning the error terms
@xmath42 we impose the following assumptions .
* @xmath42 is a strictly stationary series with mean zero and @xmath43 with the following properties : we have @xmath44 and both @xmath42 and @xmath45 satisfy a functional central limit theorem , i.e. , @xmath46 and @xmath47 as @xmath48 , for constants @xmath49 . here
@xmath50 and @xmath51 denote ( standard ) brownian motions ( wiener processes ) with start in @xmath52 .
* @xmath42 is a strong mixing strictly stationary times series with @xmath53 for some @xmath54 , and with mixing coefficients , @xmath55 , satisfying @xmath56 in assumption ( e1 ) and the rest of the paper @xmath57 denotes weak convergence in the space @xmath58 $ ] of all cadlag functions equipped with the skorokhod metric @xmath59 .
the assumption that @xmath42 satisfies an invariance principle can be regarded as a nonparametric definition of the @xmath60 property ensuring that the partial sums converge weakly to a ( scaled ) brownian motion @xmath50 . for
a parametrically oriented definition see stock ( 1994 ) .
particularly , the scale parameter @xmath61 is given by @xmath62 also introduce the notations @xmath63 if the @xmath64 are uncorrelated , we have @xmath65 , and @xmath66 . as a non - trivial example for processes satisfying ( e1 )
let us consider arch processes .
a time series @xmath67 satisfies arch(@xmath68 ) equations , if there exists a sequence of i.i.d .
non - negative random variables , @xmath69 , such that @xmath70 where @xmath71 , @xmath72 , @xmath73 this model is often applied to model conditional heteroscedasticity of an uncorrelated sequence @xmath42 with @xmath74 for all @xmath75 , by putting @xmath76 .
a common choice for @xmath77 is to assume that the @xmath77 are i.i.d . with common standard normal distribution . in giraitis
et al . ( 2003 ) it has been shown that an unique and strictly stationary solution exists and satisfies @xmath78 , if @xmath79 in addition , under these conditions the functional central limit theorem ( [ fclt2 ] ) holds . the rate of decay of the coefficients @xmath80 controls the asymptotic behavior of @xmath81 . if for some @xmath82 and @xmath83 we have @xmath84 , @xmath85 , then there exists @xmath86 such that @xmath87 for @xmath88 .
thus , depending on the rate of decay ( e2 ) may also holds .
assumption ( e2 ) will be used to verify a tightness criterion . combined with appropriate moment conditions it implies the invariance principles ( [ fclt1 ] ) and ( [ fclt2 ] ) .
we will now introduce the class of dickey - fuller processes and related detection procedures . recall that the least squares estimator of the parameter @xmath89 in model ( [ armodel ] ) is given by @xmath90 to test the null hypothesis @xmath91 , one forms the dickey - fuller ( df ) test statistic @xmath92 suppose at this point that the @xmath64 are uncorrelated . provided @xmath93 ,
@xmath94 as @xmath48
. however , @xmath95 has a different convergence rate and a non - normal limit distribution , if @xmath96 .
it is known that @xmath97 as @xmath48 , see white ( 1958 ) , fuller ( 1976 ) , rao ( 1978 , 1980 ) , dickey and fuller ( 1979 ) , and evans and savin ( 1981 ) . recall that @xmath50 denotes standard brownian motion .
based on that result one can construct a statistical level @xmath0 test , which rejects the null hypothesis @xmath91 of a unit root against the alternative @xmath98 if @xmath99 , where the critical value @xmath14 is the @xmath0-quantile of the distribution of @xmath100 . more generally , we want to construct a detection rule which provides a signal if there is some change - point @xmath101 such that @xmath102 form a random walk ( unit root process ) , and @xmath103 form an @xmath104 with dependent innovations .
this means , the alternative hypothesis is @xmath105 , where @xmath106 , @xmath107 , specifies that @xmath108 where @xmath109 .
however , for the calculation of the detection rule to be introduced now knowledge of a specific alternative hypothesis is not required .
a naive approach to monitor a time series to check for deviations from the unit root hypothesis is to apply the df statistic at each time point using the most recent observations .
a more sophisticated version of this idea is to modify the df statistic to ensure that summands in the numerator have small weight if their time distance to the current time point is large . to define such a detection rule ,
let us introduce the following sequential kernel - weighted dickey - fuller ( df ) process @xmath110,\ ] ] where @xmath111 . here and in the following we put @xmath112 for convenience .
note that @xmath113 plays the role of the current time point .
the non - negative smoothing kernel @xmath114 is used to attach smaller weights to summands from the distant past to avoid that such summands dominate the sum .
thus , kernels ensuring that @xmath115 , @xmath116 , is decreasing are appropriate , but that property is not required .
we do not use kernel weights in the denominator , since it is used to estimate a nuisance parameter .
we will require the following regularity conditions for @xmath117 .
* @xmath118 , @xmath119 and @xmath120 .
* @xmath121 is @xmath122 with bounded derivative .
* @xmath121 has bounded variation .
note that it is not required to use a kernel with compact support .
the parameter @xmath123 is used as a scaling constant in the kernel and defines the memory of the procedure .
for instance , if @xmath124 if @xmath125 $ ] and @xmath126 otherwise , the process @xmath127 looks back @xmath128 observations .
we will assume that @xmath129 for some @xmath130 .
that condition ensures that the number of observations used by @xmath131 gets larger as @xmath132 increases . note that the parameter @xmath133 , which will also appear in the limit distributions , could be absorbed into the kernel @xmath114 .
however , in practice one usually fixes a kernel @xmath114 and chooses a bandwidth @xmath128 relative to the time horizon @xmath2 .
( [ asszeta ] ) is therefore not restrictive . since small values of @xmath134
provide evidence for the alternative that the time series is stationary , intuition suggests that the control chart should give a signal if @xmath131 is smaller than a specified control limit @xmath135 .
hence , we define @xmath136 we will assume that the start of monitoring , @xmath137 , is given by @xmath138 a reasonable approach to choose @xmath14 is to control the type i error rate @xmath139 , i.e. , to ensure that @xmath140 where @xmath141 indicates that the probability is calculated assuming that @xmath16 is a random walk corresponding to the null hypothesis @xmath91 . in the next section we will show that @xmath131 converges weakly to some stochastic process @xmath142 depending on the nuisance parameter @xmath143 and that @xmath144 converges in distribution to @xmath145 : { \mathcal{d}}_\vartheta(s ) < c \ } $ ] . hence ,
if @xmath14 is chosen from the asymptotic distribution via ( [ aslimitalpha ] ) , @xmath146 is a function of @xmath147 .
therefore , the basic idea is to estimate @xmath147 at each time point using only past and current data , and to use the corresponding limit .
our estimator for @xmath147 will be based on a newey - west type estimator , thus circumventing the problem to specify the short memory dynamics of the process explicitly .
let @xmath148 and denote by @xmath149 , @xmath150 , the autocorrelation function of the time series @xmath42 . since @xmath151 if @xmath96 , we can estimate @xmath152 and @xmath153 under the null hypothesis by @xmath154 the parameter @xmath155 can now be estimated by the newey - west estimator given by @xmath156 where @xmath157 are the bartlett weights and @xmath158 is a lag truncation parameter , see newey and west ( 1987 ) . andrews ( 1991 ) studies more general weighting functions and shows that the rate @xmath159 is sufficient for consistency .
the dickey - fuller control chart for correlated time series works now as follows . at each time point
@xmath160 we estimate @xmath147 by @xmath161 and calculate the corresponding estimated control limit @xmath162 .
a signal is given if @xmath131 is less than the estimated control limit , i.e. , we use the rule @xmath163 alternatively , one may use a transformation of @xmath131 , namely @xmath164.\ ] ] it seems that this transformation idea dates back to phillips ( 1987 ) .
we will show that for arbitrary @xmath147 the process @xmath165 converges weakly to the limit of @xmath131 for @xmath166 .
consequently , if @xmath14 denotes the control limit ensuring that @xmath167 has size @xmath0 when @xmath166 , then the detection rule @xmath168 has asymptotic size @xmath0 for any @xmath147 . in the next section we shall show that both procedures are asymptotically valid .
inference on the ar parameter in the unit root case is often based on the @xmath160-statistic associated with @xmath131 , which gives rise to dickey - fuller @xmath75-processes .
the dickey - fuller @xmath160-statistic , @xmath169 , associated with @xmath170 , is the standard computer output quantity when running a regression of @xmath171 on @xmath172 . for a sample @xmath173 , the statistic @xmath169 is defined as @xmath174 where @xmath175 with @xmath176 .
the formula for @xmath169 motivates to scale @xmath131 analogously .
hence , let us define the weighted @xmath75-type df process by @xmath177,\ ] ] and @xmath178 .
@xmath179 is a weighted version of @xmath169 calculated using the observations @xmath180 , and attaching kernel weights @xmath181 to the @xmath182th summand in the numerator .
the associated detection rule for known @xmath147 is defined as @xmath183 with @xmath184 such that @xmath185 .
again , it turns out that the asymptotic limit of @xmath186 depends on the nuisance parameter @xmath147 .
the weighted @xmath75-type df control chart with estimated control limits is defined as @xmath187 alternatively , one can transform the process to achieve that the asymptotic limit is invariant with respect to @xmath147 .
we define @xmath188.\ ] ] we will show that the detection rule @xmath189 has asymptotic type i error equal to @xmath0 for all @xmath147 .
in this section we provide functional central limit theorems for the dickey - fuller processes defined in the previous section under a random walk model assumption corresponding to the null hypothesis @xmath190 in model ( [ armodel ] ) , and the related central limit theorem for the associated stopping rules .
these results can be used to design tests and detection procedures having well - defined statistical properties under the null hypothesis .
we start with the following functional central limit theorem providing the limit distribution of the weighted df process @xmath191 , @xmath192
$ ] , which extends phillips ( 1987 , th . 3.1 c ) .
[ thbasic ] assume the time series @xmath193 satisfies model ( [ armodel ] ) with @xmath96 such that ( e1 ) and ( k1)-(k3 ) hold
. then @xmath194,d)$},\ ] ] as @xmath48 , where the stochastic process @xmath195 @xmath196 $ ] , @xmath197 , is continuous w.p . @xmath26 .
note that the asymptotic limit is distribution - free if and only if @xmath198 which holds if the error terms are uncorrelated .
otherwise , the distribution of @xmath142 depends sensitively on @xmath147 . if @xmath96 we have @xmath151 and @xmath199 for all @xmath160 .
this yields the representation @xmath200,\ ] ] where the @xmath58 $ ] -valued stochastic processes @xmath201 , @xmath202 , and @xmath203 are given by @xmath204 for @xmath196 $ ] .
let us first show that @xmath205 } | \widetilde{r}_t(s ) - \mu(s ) |
\stackrel{p}{\to } 0,\ ] ] as @xmath48 , where @xmath206.\ ] ] consider @xmath207 ( [ asszeta ] ) ensures that @xmath208 } \max_i | ( { { \lfloor ts \rfloor } } - i)/h - \zeta(s - i / t ) | = o(1 ) $ ] yielding @xmath209 uniformly in @xmath210 $ ] , because @xmath114 is lipschitz continuous and of bounded variation , cf .
theorem 3.3(ii ) of steland ( 2004 ) .
it remains to estimate @xmath211 .
the assumptions on @xmath42 ensure that @xmath212 as @xmath48 , where @xmath213 .
hence , eventually for equivalent versions , we may assume that @xmath214 a.s .
, for @xmath48
. by ( k3 ) the stieltjes integrals @xmath215 and @xmath216 are well defined ( via integration by parts ) , and @xmath217 } \left| \int_0^s k ( \zeta(s - r ) ) \ , d z_t(r ) - \rho \int_0^s k ( \zeta(s - r ) ) \ , d b^{(2)}(r ) \right| \stackrel{a.s.}{\to } 0,\ ] ] as @xmath48 . obviously , @xmath218 } | \widetilde{r}_t(s ) - e \widetilde{r}_t(s ) | = \sup_{s \in [ \kappa,1 ] } \frac{\sqrt{t } } { { { \lfloor ts \rfloor } } } \left| \int_0^s k ( ( { { \lfloor ts \rfloor } } - { { \lfloor tr \rfloor } } ) /h ) d z_t(r ) \right| \\ & & \le \sup_{s \in [ \kappa,1 ] } \frac { \rho \sqrt{t } } { { { \lfloor ts \rfloor } } } \left| b^{(2)}(r ) k\bigl [ \frac { { { \lfloor ts \rfloor } } - { { \lfloor tr \rfloor } } } { h } \bigr ] \bigr|_{r=0}^{r = s } - \int_0^s b^{(2)}(r ) k ( ( { { \lfloor ts \rfloor } } - { { \lfloor t(dr ) \rfloor } } ) /h ) \right| \\ & & \quad + \sup_{s \in [ \kappa,1 ] } \frac { \sqrt{t } } { { { \lfloor ts \rfloor } } } \biggl| [ z_t(r ) - \rho b^{(2)}(r ) ] k ( ( { { \lfloor ts \rfloor } } - { { \lfloor tr \rfloor } } ) /h ) \bigr|_{r=0}^{r = s } \\ & & \qquad \quad - \int_0^s [ z_t(r ) - \rho b^{(2)}(r ) ] k ( ( { { \lfloor ts \rfloor } } - { { \lfloor t(dr ) \rfloor } } ) / h ) \biggr|.\end{aligned}\ ] ] noting that the total variation of the functions @xmath219/h ) $ ] , @xmath210 $ ] , @xmath220 , is uniformly bounded , the right side of the above display can be estimated by @xmath221 therefore , ( [ rem1 ] ) holds true .
let us now consider @xmath222 .
we will first show that , up to terms of order @xmath223 , @xmath201 is a functional of @xmath224.\ ] ] again , under the assumptions of the theorem , @xmath127 converges weakly to @xmath225 , where @xmath50 denotes brownian motion and @xmath226 is a constant . for brevity of notation let @xmath227.\ ] ] integration by parts yields @xmath228 due to ( k2 ) the @xmath229 term is uniform in @xmath210 $ ] .
next note that @xmath230 we are now in a position to verify joint weak convergence of numerator and denominator of @xmath131 .
the lipschitz continuity of @xmath114 ensures that up to terms of order @xmath229 for all @xmath231 the linear combination @xmath232 is a functional of @xmath127 , and that functional is continuous .
therefore , the continuous mapping theorem ( cmt ) entails weak convergence to the stochastic process @xmath233 \\ & & \quad + \lambda_2 \frac{\eta^2}{s^2 } \int_0^s b(r)^2 \ , dr.\end{aligned}\ ] ] this verifies joint weak convergence of @xmath234 .
hence , the result follows by the cmt .
( k2 ) also ensures that @xmath235 $ ] w.p . @xmath26 . the central limit theorem ( clt ) for the detection procedure @xmath167 , which requires knowledge of @xmath147 , appears as a corollary .
[ corst ] under the assumptions of theorem [ thbasic ] we have for any control limit @xmath236 @xmath237 : { \mathcal{d}}_\vartheta(s ) < c \}\ ] ] as @xmath48 , where @xmath238 is defined in ( [ defd ] ) .
observe that by definition of @xmath167 @xmath239 } d_t(s ) \ge c \leftrightarrow \sup_{s \in [ \kappa , x ] } -d_t(s ) \le -c\ ] ] for any @xmath240 .
hence it suffices to show that @xmath241 } -d_t(s ) \le -c ) \to p ( \sup_{s \in [ \kappa , x ] } -{\mathcal{d}}_\vartheta(s ) \le -c ) , \ ] ] where @xmath142 denotes the limit process given in theorem [ thbasic ] .
using the skorokhod / dudley/ wichura representation theorem and a result due to lifshits ( 1982 ) , this fact can be shown along the lines of the proof of theorem 4.1 in steland ( 2004 ) , if @xmath236 , since @xmath235 $ ] a.s .
for brevity we omit the details .
let us now show consistency of the detection procedure @xmath242 , which uses estimated control limits .
[ thestcl ] assume ( e1 ) and ( e2 ) , ( k1)-(k3 ) , and in addition that the lag truncation parameter , @xmath158 , of the newey - west estimator satisfies @xmath243 then the weighted dickey - fuller type control chart with estimated control limit , @xmath244 , is consistent , i.e. , @xmath245 as @xmath48 . note
that the equivalence @xmath246 } d_t(s ) / c ( \widehat{\vartheta } _ { { { \lfloor ts \rfloor } } } ) \ge 1 $ ] implies @xmath247 } \frac{d_t(s)}{c ( \widehat{\vartheta } _ { { { \lfloor ts \rfloor } } } ) } < 1 \right).\ ] ] let us first show that the function @xmath248 is continuous . note that the process @xmath238 can be written as @xmath249 for a.s .
continuous processes @xmath250 and @xmath251 not depending on @xmath147 , where particularly @xmath252 and @xmath253.\ ] ] let @xmath254 be a sequence with @xmath255 , as @xmath256 . clearly , for each @xmath257 of the underlying probability space with @xmath258 , we have @xmath259 @xmath256
. hence , @xmath208 } { \mathcal{d}}_{\vartheta_n}(s ) \stackrel{d}{\to } \sup_{s \in [ \kappa,1 ] } { \mathcal{d}}_{\vartheta^*}(s ) $ ] , as @xmath256 .
since @xmath208 } { \mathcal{d}}_{\vartheta^*}(s ) $ ] has a continuous density , this is equivalent to pointwise convergence of the d.f . @xmath260 } { \mathcal{d}}_{\vartheta_n}(s ) \le z ) $ ] to @xmath261 } { \mathcal{d}}_{\vartheta^*}(s ) \le z ) $ ] , as @xmath256 , for all @xmath116 .
hence , @xmath262 as @xmath256
. next we show @xmath263 as @xmath48 , in @xmath264 $ ] . since for each @xmath210 $ ] we have @xmath265 , for @xmath48 ,
fidi convergence follows immediately .
it remains to verify tightness .
recall the definitions ( [ defnw1 ] ) and ( [ defnw2 ] ) and that @xmath266 under @xmath267 . fix @xmath268 and consider the process @xmath269 ,
@xmath210 $ ] , which is a functional of @xmath270 . clearly , by the cauchy - schwarz inequality and ( e1 )
@xmath271 for some @xmath54 .
further , since @xmath272 and @xmath273 , the mixing coefficients @xmath274 of @xmath275 satisfy @xmath276 where @xmath277 are the mixing coefficients of @xmath42 . due to ( e1 ) we can apply yokohama ( 1980 , th.1 ) with @xmath278 to conclude that for @xmath279 @xmath280 now the decomposition @xmath281 and the triangle inequality yield @xmath282 since , firstly , we may assume @xmath283 , and , secondly , both @xmath284 and @xmath285 are bounded away from @xmath52 and @xmath286 for @xmath287 . consequently , @xmath288 and therefore vaart and wellner ( 1986 , ex .
2.2.3 ) implies tightness of the process @xmath289 \ } $ ] for fixed @xmath290 .
note that @xmath291 . by the triangle inequality we have @xmath292
yielding @xmath293 hence , @xmath294 \ } $ ] is tight in the product space , which implies weak convergence of @xmath295 \ } $ ] to @xmath296 . the final step is to verify @xmath297 } d_t(s ) / c ( \widehat{\vartheta } _ { { { \lfloor ts \rfloor } } } ) \stackrel{d}{\to } \inf_{s \in [ \kappa,1 ] } { \mathcal{d}}_\vartheta(s ) / c ( \vartheta ) , \ ] ] as @xmath48 , since this implies that ( [ stern ] ) converges to @xmath298 } { \mathcal{d}}_\vartheta(s ) < c(\vartheta ) ) = \alpha $ ] , as @xmath48 . due to
( [ wctheta ] ) we can conclude that @xmath299 in the product space @xmath300 ) ^2 $ ] .
note that the mapping @xmath301,d ) ^2 \to ( { \mathbb{r } } , { \mathcal{b } } ) $ ] given by @xmath302 } \frac{x(s)}{c(y(s ) ) } , \qquad x , y \in d[\kappa,1 ] , \ y
\in { \mathbb{r } } , \ ] ] is continuous in all @xmath303)^2 $ ] . since @xmath304 $ ] w.p . @xmath26 and @xmath305 , ( [ wc2 ] ) follows .
it remains to provide the related weak convergence results for the transformed process @xmath165 and its natural detection rule @xmath306 .
assume ( e1),(e2 ) , and ( k1)-(k3 ) .
additionally assume that the lag truncation parameter , @xmath158 , of the newey - west estimator satisfies @xmath243 then , @xmath307,d)$}\ ] ] as @xmath48 , and for the transformed dickey - fuller type control chart we have @xmath308 as @xmath48 .
particularly , the asymptotic distributions are invariant with respect to @xmath147 . as shown above , @xmath309 as @xmath48 , which implies that @xmath310 if @xmath48 , yielding @xmath311 let us now derive ( functional ) central limit theorems for the weighted dickey - fuller @xmath75-processes and the associated detection rules .
we start with the process @xmath186 under the random walk null hypothesis .
[ thttype ] assume ( e1 ) , and ( k1)-(k3 )
. then @xmath312,d)$}\ ] ] as @xmath48 , where @xmath313 for @xmath196 $ ] and @xmath314 . here
@xmath316 is continuous a.s .
note that again the limit depends on the nuisance parameter @xmath147 and is distribution - free if and only if @xmath166 . by definition @xmath317 where @xmath318 with @xmath319 for @xmath196 $ ] .
note that for @xmath320 @xmath321 hence , we obtain @xmath322 from the proof of theorem [ thbasic ] we know that @xmath323 } { { \lfloor ts \rfloor } } ^{-2 } \sum_{t=1}^ { { { \lfloor ts \rfloor } } } y_{t-1}^2 = \sup_{s \in ( 0,1 ] } \left ( \frac { t } { { { \lfloor ts \rfloor } } } \right)^2 \int_0^s ( t^{-1/2 } y _ { { { \lfloor tr \rfloor } } } ) ^2 \ , dr = o_p(1)\ ] ] and @xmath323 } \left| { { \lfloor ts \rfloor } } ^{-1 } \sum_{t=1}^ { { { \lfloor ts \rfloor } } } \epsilon_t y_{t-1 } \right| \le \sup_{s \in ( 0,1 ] } \left| { { \lfloor ts \rfloor } } ^{-1/2 } \sum_{t=1}^ { { { \lfloor ts \rfloor } } } \epsilon_t \right| \sup_{s \in ( 0,1 ] } | { { \lfloor ts \rfloor } } ^{-1/2 } y _ { { { \lfloor ts \rfloor } } } | = o_p(1).\ ] ] combining these facts with @xmath324 } { { \lfloor ts \rfloor } } | \widehat{\rho } _ { { { \lfloor ts \rfloor } } } - 1 | = o_p(1 ) $ ] , we obtain @xmath325 where the @xmath229 term is uniform in @xmath196 $ ] . because ( e1 ) implies that @xmath326 we may apply the law of large numbers for time series ( brockwell and davis ( 1991 ) , th .
7.1.1 ) and obtain , since stochastic convergence to a constant yields stochastic convergence in the skorokhod topology , @xmath327 as @xmath48
. we shall now show joint weak convergence of @xmath328 , @xmath196 $ ] .
let @xmath329 and consider @xmath330.\ ] ] the proof of theorem [ thbasic ] implies that @xmath331 as @xmath48 . due to ( [ convconst ] ) , we obtain @xmath332 as @xmath48 .
therefore , the cmt implies that @xmath333 and @xmath334 as @xmath48 , yielding the assertion .
we are now in the position to establish consistency of the @xmath75-type detection rule @xmath335 which uses estimated control limits .
notice that theorem [ thttype ] implies that @xmath336 is given by @xmath337 } \widetilde{{\mathcal{d}}}_\vartheta(s ) < c(\vartheta ) ) = \alpha $ ] .
assume ( e1),(e2 ) , ( k1)-(k3 ) , and additionally that the lag truncation parameter of the newey - west estimator satisfies @xmath243 then the @xmath75-type weighted dickey - fuller control chart with estimated control limits , @xmath338 , is consistent , i.e. , @xmath339 as @xmath48 . the result is shown along the lines of the proof of theorem [ thestcl ] , since the process @xmath316 is continuous w.p .
@xmath26 , and is a continuous function of @xmath147 . finally , for the transformed process @xmath340 and the associated control chart
@xmath341 we have the following result .
assume ( e1),(e2 ) , ( k1)-(k3 ) , and @xmath243 then the transformed @xmath75-type weighted df process @xmath340 , defined in ( [ deftranst ] ) , converges weakly , @xmath342,d)$},\ ] ] as @xmath48 , and for the transformed @xmath75-type weighted df control chart we have @xmath343 particularly , the asymptotic distribution is invariant with respect to @xmath147 .
note that the first term of @xmath340 converges weakly to @xmath344 , which has the form @xmath345/ [ \int_0^s b^2(r ) \ , dr ] ^{1/2 } $ ] .
hence , the construction of the correction term is as for @xmath165 .
in econometric applications , the stationary alternatives of interest are often of the form @xmath346 with @xmath347 small .
to mimic this situation asymptotically , we consider a local - to - unity model where the ar parameter depends on @xmath132 and tends to @xmath348 , as the time horizon @xmath132 increases . the functional central limit theorem given below shows that the asymptotic distribution under local - to - unity alternatives is also affected by the nuisance parameter @xmath147 .
however , the term which depends on the parameter parameterising the local alternative does not depend on @xmath147 ( or @xmath349 ) . therefore ,
if one takes the nuisance parameter @xmath147 into account when designing a detection procedure , we obtain local asymptotic power .
let us assume that we are given an array @xmath350 of observations satisfying @xmath351 where the sequence of ar parameters @xmath352 is given by @xmath353 for some constant @xmath354 .
@xmath42 is a mean - zero stationary i(0 ) process satisfying ( e1 ) . for brevity of notation @xmath355
denotes in this section the process ( [ defdt ] ) with @xmath171 replaced by @xmath356
. the limit distribution will be driven by an ornstein - uhlenbeck process .
recall that the ornstein - uhlenbeck process @xmath357 with parameter @xmath358 is defined by @xmath359,\ ] ] where @xmath360 denotes brownian motion .
[ thlocalunity ] assume ( e1 ) , and ( k1)-(k3 ) . under the local - to - unity model ( [ localtounity ] ) we have for the weighted dickey - fuller process @xmath361 as @xmath48 , where the a.s .
@xmath362 $ ] -valued process @xmath363 is given by @xmath364 for @xmath196 $ ] , and @xmath365 . here
@xmath366 denotes the ornstein - uhlenbeck process defined in ( [ defou ] ) .
further , @xmath367 : { \mathcal{d}}_\vartheta^a(s ) < c \ } , \qquad \mbox{as $ t \to \infty $ } .\ ] ] the crucial arguments to obtain _ joint _ weak convergence of numerator and denominator of @xmath127 have been given in detail in the proof of theorem [ thbasic ] .
therefore , we give only a sketch of the proof stressing the essential differences .
first , note that @xmath368 for the step function @xmath369 , @xmath370 $ ] , which has uniformly bounded variation and converges uniformly in @xmath371 to the exponential @xmath372 .
hence , firstly , the stochastic stieltjes integral @xmath373 exists ( via integration by parts ) , and , secondly , by estimating the terms of the decomposition @xmath374 we see that @xmath375 as @xmath48 .
next , note that in the local - to - unity model we have @xmath376 for all @xmath377 .
this yields the decomposition @xmath378 where for @xmath196 $ ] @xmath379 the term @xmath380 can be treated as in the proof of theorem [ thbasic ] , namely , @xmath381 from the proof of theorem [ thbasic ] we know that due to ( e1 ) @xmath382 as @xmath48 .
consider now @xmath383 . by definition of @xmath384
we obtain @xmath385 where due to ( k2 ) the @xmath229 term is uniform in @xmath196 $ ] .
hence , @xmath380 , @xmath383 , and @xmath203 are functionals of @xmath127 up to terms of order @xmath229 .
consequently , joint weak convergence of @xmath386 can be shown along the lines of the proof of theorem [ thbasic ] , and the cmt yields the result .
to investigate the statistical properties of the proposed monitoring procedure we performed a simulation study .
we used the following arma(1,1 ) simulation model .
suppose @xmath387 where @xmath388 , @xmath389 is a sequence of independent @xmath390-distributed error terms , and @xmath89 and @xmath391 are parameters .
we investigated the cases given by @xmath392 and @xmath393 .
clearly , @xmath394 corresponds to the unit root null hypothesis . for @xmath395
the innovation terms are uncorrelated corresponding to @xmath396 .
this simulation model was also used in steland ( 2006 ) , where a monitoring procedure based on the kpss unit root test is studied in detail . since part of the parameter settings used below
are identical , the results of the present numerical study can be compared with the corresponding results in steland ( 2006 ) . to study the monitoring rules with estimated control limits critical values for a significance level of @xmath397 were taken from the limit process defined in ( [ defd ] ) with estimated nuisance parameter . to down - weight past contributions a gaussian kernel with bandwidth @xmath398 was used .
the nuisance parameter @xmath147 was estimated by the newey - west estimator at time point @xmath75 with lag truncation parameter @xmath158 chosen by @xmath399 , @xmath400 .
the start of monitoring , @xmath137 , affects the properties and has to be chosen carefully . for the rule @xmath244 we used @xmath401 , whereas for @xmath402 a larger value , @xmath403 , yielded better results . to investigate the properties of the monitoring rule
, we estimate empirical rejection rates of the test which rejects the unit root null hypothesis if the procedure gives a signal , the average delay , and the average conditional delay given a signal . for the detection rule @xmath244 the arl
is defined by @xmath404 .
we define the carl as @xmath405 .
the definitions for @xmath402 are analogous .
note that the conditional delay is very informative under the alternative , since it informs us how quick the method reacts if it reacts at all . in the tables
average delays are given in brackets and conditional delay in parentheses .
table [ tab1 ] provides the results for the monitoring procedures @xmath244 and @xmath402 using estimated control limits .
the curves @xmath184 were obtained by simulating from the limit laws .
overall , @xmath406 performed well .
the performance of the @xmath75-type procedure is disappointing .
when inspecting the carl values , the results seem to be mysterious .
e.g. when comparing the carl for @xmath407 and @xmath408 if @xmath409 , the procedure seems to misbehave . to explore the reason , figure [ fig1 ] provides a part of the distribution of @xmath410 .
it can be seen that the percentage of simulated trajectories leading to immediate detection increases considerably , but the contribution of these cases to the calculation of the carl is negligible .
the other trajectories yielding a signal are hard to detect , and the signals are spread over the remaining time points with many late signals , which suffice to yield large carl values .
this fact shows that a single number as the carl can not summarized the statistical behavior sufficiently .
it highlights the benefit that the random walk null hypothesis can often be rejected very early .
the simulation results for the control charts using transformed statistics are summarized in table [ tab2 ] .
here we used exact control limits obtained by simulation using 20,000 repetitions .
comparing the transformation control statistics with these control limits yields quite accurate results if @xmath411 .
the @xmath75-type version is preferable for @xmath412 . comparing the methods @xmath244 ( using estimated control limits ) and @xmath413 ( using transformed statistics ) ,
our results indicate that the more computer - intensive approach to use estimated control limits provides more accurate results .
cccccc @xmath414 & + & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath419 + + + @xmath26 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 + @xmath425 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 + & @xmath431 & @xmath432 & @xmath433 & @xmath434 & @xmath435 + & @xmath436 $ ] & @xmath437 $ ] & @xmath438 $ ] & @xmath439 $ ] & @xmath440 $ ] + @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath418 & @xmath445 + & @xmath446 & @xmath447 & @xmath448 & @xmath449 & @xmath450 + & @xmath451 $ ] & @xmath452 $ ] & @xmath453 $ ] & @xmath454 $ ] & @xmath455 $ ] + @xmath456 & @xmath457 & @xmath458 & @xmath459 & @xmath460 & @xmath26 + & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 + & @xmath466 $ ] & @xmath439 $ ] & @xmath467 $ ] & @xmath468 $ ] & @xmath469 $ ] + + + @xmath26 & @xmath470 & @xmath471 & @xmath472 & @xmath473 & @xmath474 + @xmath425 & @xmath475 & @xmath476 & @xmath477 & @xmath478 & @xmath479 + & @xmath480 & @xmath481 & @xmath482 & @xmath483 & @xmath484 + & @xmath485 $ ] & @xmath486 $ ] & @xmath487 $ ] & @xmath488 $ ] & @xmath489 $ ] + @xmath441 & @xmath490 & @xmath420 & @xmath491 & @xmath430 & @xmath26 + & @xmath492 & @xmath493 & @xmath493 & @xmath494 & @xmath495 + & @xmath496 $ ] & @xmath497 $ ] & @xmath498 $ ] & @xmath499 $ ] & @xmath500 $ ] + @xmath456 & @xmath501 & @xmath502 & @xmath503 & @xmath504 & @xmath26 + & @xmath505 & @xmath506 & @xmath492 & @xmath507 & @xmath508 + & @xmath509 $ ] & @xmath510 $ ] & @xmath511 $ ] & @xmath512 $ ] & @xmath513 $ ] + for @xmath514 ( circles ) and @xmath408 ( crosses).,width=302 ] cccccc @xmath414 & + & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath419 + + + @xmath26 & @xmath515 & @xmath515 & @xmath516 & @xmath517 & @xmath518 + @xmath425 & @xmath519 & @xmath516 & @xmath520 & @xmath521 & @xmath522 + & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath527 + & @xmath528 $ ] & @xmath529 $ ] & @xmath530 $ ] & @xmath531 $ ] & @xmath532 $ ] + @xmath441 & @xmath533 & @xmath534 & @xmath535 & @xmath536 & @xmath537 + & @xmath538 & @xmath538 & @xmath539 & @xmath461 & @xmath540 + & @xmath541 $ ] & @xmath542 $ ] & @xmath543 $ ] & @xmath544 $ ] & @xmath545 $ ] + @xmath456 & @xmath546 & @xmath547 & @xmath548 & @xmath549 & @xmath26 + & @xmath550 & @xmath551 & @xmath552 & @xmath550 & @xmath553 + & @xmath554 $ ] & @xmath555 $ ] & @xmath556 $ ] & @xmath557 $ ] & @xmath558 $ ] + + + @xmath26 & @xmath559 & @xmath560 & @xmath561 & @xmath562 & @xmath563 + @xmath425 & @xmath564 & @xmath565 & @xmath566 & @xmath567 & @xmath568 + & @xmath569 & @xmath570 & @xmath494 & @xmath538 & @xmath571 + & @xmath572 $ ] & @xmath573 $ ] & @xmath574 $ ] & @xmath575 $ ] & @xmath576 $ ] + @xmath441 & @xmath577 & @xmath578 & @xmath579 & @xmath580 & @xmath26 + & @xmath581 & @xmath582 & @xmath583 & @xmath584 & @xmath585 + & @xmath586 $ ] & @xmath587 $ ] & @xmath588 $ ] & @xmath589 $ ] & @xmath590 $ ] + @xmath456 & @xmath591 & @xmath592 & @xmath593 & @xmath594 & @xmath26 + & @xmath595 & @xmath596 & @xmath492 & @xmath597 & @xmath508 + & @xmath598 $ ] & @xmath599 $ ] & @xmath600 $ ] & @xmath601 $ ] & @xmath513 $ ] +
the support of deutsche forschungsgemeinschaft ( sfb 475 , _ reduction of complexity in multivariate data structures _ ) is gratefully acknowledged .
i thank dipl .- math .
sabine teller for proof - reading .
kwiatkowski , d. , phillips , p.c.b . ,
schmidt , p. , and shin , y. ( 1992 ) .
testing the null hypothesis of stationary against the alternative of a unit root : how sure are we that economic time series have a unit root ?
_ journal of econometrics _ , * 54 * , 159 - 178 . | aiming at monitoring a time series to detect stationarity as soon as possible , we introduce monitoring procedures based on kernel - weighted sequential dickey - fuller ( df ) processes , and related stopping times , which may be called weighted dickey - fuller control charts . under rather weak assumptions , ( functional ) central limit theorems are established under the unit root null hypothesis and local - to - unity alternatives . for general dependent and heterogeneous innovation sequences
the limit processes depend on a nuisance parameter . in this case of practical interest
, one can use estimated control limits obtained from the estimated asymptotic law .
another easy - to - use approach is to transform the df processes to obtain limit laws which are invariant with respect to the nuisance parameter .
we provide asymptotic theory for both approaches and compare their statistical behavior in finite samples by simulation . +
* keywords : * autoregressive unit root , change point , control chart , nonparametric smoothing , sequential analysis , robustness . weighted dickey - fuller processes for detecting stationarity ansgar steland + institute of statistics , + rwth aachen university , germany + [email protected] |
the characteristics of the galaxy populations in the cores of rich clusters appear to vary strongly as a function of redshift , with an increasing fraction of blue galaxies in clusters at @xmath3 ( the butcher - oemler effect , butcher & oemler 1984 ) .
the search for the processes responsible for these rapid changes has become a vigorous field of research and at least for the blue cluster galaxies a broad concensus has been reached about the nature of this population .
the bulk are star - forming cluster members , they cover a wide range of luminosities , from a few @xmath4 downwards , and have emission line strengths which indicate moderate to high star formation rates ( abraham et al .
1996 ; balogh et al . 1999 ; barger et al .
1996 ; couch & sharples 1987 ; dressler & gunn 1992 ; dressler et al . 1999 , d99 ; fabricant et al .
1994 ; fisher et al . 1998 ; lavery & henry 1988 ) .
the optical morphologies of these galaxies were investigated firstly with high - resolution ground - based ( lavery et al .
1992 ) and more recently with _ hubble space telescope _
( _ hst _ ) imaging ( dressler et al .
1994a ; couch et al .
1994 , 1998 ; oemler et al . 1997
; smail et al . 1997 , s97 ; fabricant et al .
these studies have shown that most of the blue star - forming galaxies have strong disk components and a large fraction also appear to be disturbed , with structures suggesting mergers or tidal interactions .
thus the increasing proportion of blue galaxies within distant clusters is associated with an increasing number of actively star - forming disk galaxies in these environments .
the large fraction of apparently disturbed galaxies suggests that galaxy - galaxy interactions or the effects of the cluster tidal field on these disk galaxies could be responsible for the butcher - oemler effect ( moore , lake & katz 1998 ) .
the spectroscopic studies of distant clusters uncovered another population of luminous galaxies which is not seen in similar numbers in local clusters .
these are galaxies with strong balmer absorption ( typically measured using the h@xmath5@xmath64104 line and coming predominantly from a stars ) and no [ oii]@xmath63727 emission .
the large population of a stars is a sign that the galaxy was actively forming massive stars in the recent past ( @xmath7gyr ) , while the lack of emission lines suggests that this star formation has now ceased .
together the spectral properties are interpreted as showing that the galaxy is in a post - starburst phase .
almost 20% of the luminous galaxies within the cores of distant clusters fall in this post - starburst class ( e.g. d99 , although see balogh et al .
1999 ) and it has been suggested that this high fraction indicates that almost all cluster galaxies pass through this phase ( barger et al .
1996 ) .
there is an increasing realisation in the community of the importance of dust obscuration in defining the apparent properties of galaxies in the near - uv and optical , particularly at high redshifts .
this issue has been highlighted by the disagreement over the form of the star formation history of the universe as estimated from the variation in the ultraviolet luminosity density with redshift ( e.g. lilly et al .
1996 ; steidel et al .
1999 ) compared to star formation indicators at longer wavelengths such as h@xmath8 emission ( yan et al .
1999 ) or reradiated star light detected in the sub - millimeter ( e.g. blain et al . 1999 ) .
the longer wavelength tracers tend to find higher star formation densities , a result which has been attributed to dust absorption in the uv and optical ( calzetti & heckman 1999 ; meurer et al . 1997 ;
pettini et al .
1998 ) .
the issue of the effects of dust on the evolutionary cycle associated with the optical butcher - oemler effect has been disregarded by most investigations to date .
however , the recent analysis of the spectral catalog for the morphs sample ( d99 ) by poggianti et al .
( 1999 , p99 ) has uncovered evidence for dust obscuration in at least one spectral class e(a ) galaxies .
these galaxies have detectable [ oii ] emission and relatively strong balmer absorption ( ew(h@xmath9 ) .
these spectral features , along with other properties of the galaxies and of similar spectrally - classified local systems ( p99 ; poggianti & wu 1999 ) , were interpreted by p99 as probably arising from a dust - obscured starburst , with the [ oii ] line strength suppressed by dust absorption . the e(a ) class is naturally identified as the progenitor of the large population of post - starburst galaxies in the clusters . to obtain a complete view of the star formation properties of cluster galaxies free from the effects of dust obscuration we need to complement the optical studies with observations at longer wavelengths , from the near - infrared to the sub - mm and radio . in particular , observations in the sub - mm and far - infrared can trace the amount of star light absorbed by dust and reradiated at longer wavelengths .
for the most extreme starburst galaxies in the local universe this reradiated emission dominates the bolometric luminosity of the galaxies ( e.g. sanders & mirabel 1996 ) .
unfortunately , current sub - mm and far - infrared surveys lack the sensitivity to study all but the most extreme starburst galaxies ( @xmath10 yr@xmath11 ) at moderate and high redshift .
however , at radio wavelengths , the synchotron radiation from electrons accelerated by supernovae leads to a tight correlation between the radio and far - infrared fluxes of star - forming galaxies across a wide range in luminosity ( condon 1992 ) .
this means that in the absence of a radio - loud active nucleus the radio flux of a star - forming galaxy can be used to estimate its current massive star formation rate . by employing the huge collecting area of the vla we can undertake a sensitive radio survey for star formation in obscured galaxies down to limits of @xmath12@xmath13 yr@xmath11 at @xmath14 . to this end
the vla has been used in an ambitious survey to determine the evolution of the radio populations in a sample of very - rich clusters out to @xmath15 ( morrison et al .
the survey contains 14 clusters at @xmath160.18 , 12 more between @xmath170.25 and a further four at higher redshifts , @xmath180.41 ( the radio observations discussed in this paper are of the most distant cluster in this survey ) .
morrison et al . ( 1999 ) find that the population of radio galaxies with luminosities above @xmath19whz@xmath11 in the most extreme clusters has increased by a factor of @xmath206 out to @xmath15 .
they conclude that the radio galaxy population in very rich clusters has evolved significantly over the last 5gyrs .
in this paper we report on high resolution near - infrared imaging from _ hst _ and sensitive radio observations with the vla of a small field in the core of the distant cluster cl0939 + 4713 ( a851 , @xmath0 ) for which high - resolution optical _ hst _ images and extensive archival spectroscopic observations are available .
we combine these datasets to study the characteristics of galaxies in distant clusters and to search for the signatures of dust obscuration on their apparent properties . in 2
we describe the _ hst _ optical and near - infrared imaging as well as the deep vla 1.4-ghz map of this field . in
3 we then present our results on the properties of the radio - selected galaxies within our _ hst _ field drawing on published spectroscopy .
we discuss these in 4 and give our main conclusions in 5 . throughout this paper
we use @xmath21kms@xmath11mpc@xmath11 and a @xmath22 , @xmath23 cosmology . in this geometry
@xmath24 corresponds to 6.5kpc at @xmath0 .
the near - infrared _ nicmos _ observations of cl0939 + 4713 discussed here were obtained with camera 3 during the special campaign in january 1998 . due to the distortion of the _ nicmos _
dewar the wide - field camera 3 ( field - of - view 51.2@xmath25 square , sampling 0.20@xmath25 per pixel ) was not par - focal with the other instruments onboard _ hst _ and to allow the diffraction - limited operation of this camera a special campaign was organised during which the main telescope optics were refocused on the camera 3 focal plane .
the observations comprise a mosaic of four pointings through the f160w filter covering a combined field of around @xmath26 .
each pointing consists of four individual exposures of 1215s , each spatially offset by 1.1@xmath25 on a square grid to remove bad pixels and reduce the effects of flatfield variations .
the total integration time per pixel for the mosaic is thus 4.86ks .
the data were reduced in a standard manner using the _ nicmos _ data pipeline ( calnica ) and the most recent calibration files taken from the archive . any remaining features in the background
were removed with the unpedestal software kindly provided by dr .
r. van der marel which worked well , and the final mosaicing was achieved using imcombine in iraf .
the mosaic was photometrically calibrated to give magnitudes on the @xmath27 vega system .
this image provides restframe @xmath28-band morphologies at a resolution of around 1.3kpc for cluster members ( fig
. 1 ) . to create a complete catalog of objects detected within the _ nicmos _ mosaic we analysed the frame using the sextractor image analysis package ( bertin & arnoults 1996 ) .
we adopted detection criteria of a minimum area of @xmath29sq .
arcsec ( 10 pixels ) above an @xmath30 mag arcsec@xmath31 isophote and using these we detect 303 sources brighter than the approximate 3@xmath1 point - source flux limit of @xmath32 .
the vla observations of cl0939 + 4713 were obtained in b configuration at 1.4ghz between 1996 january 0608 .
the total integration time of the map is 60.7ks at an effective resolution of 5.0@xmath25 ( 33kpc at @xmath0 ) .
the map was cleaned and analysed using aips .
details of this complex procedure are given in morrison et al .
the map reaches a 1@xmath1 noise level of 9@xmath2jybeam@xmath11 .
we detect 10 sources above 27@xmath2jy ( 3@xmath1 ) within the field defined by the overlap region of our _ nicmos _ f160w mosaic and _ wide field and planetary camera 2 _ ( _ wfpc2 _ ) f702w exposures ( see fig . 1 ) .
all ten sources have bright optical / near - infrared counterparts in our _ hst _ images .
the properties of these ten sources are discussed in 3 .
the positions , fluxes and restframe radio power ( using the redshifts from table 2 and assuming an @xmath33 spectrum ) are reported in table 1 . unfortunately due to the modest resolution of the map
only two sources , # 122 and # 296 , appear to be resolved and then only marginally .
the radio and optical / near - ir images are aligned to an rms accuracy of @xmath34 using the 8 brightest radio sources in the _ nicmos _ field .
the flux limit of the radio sample corresponds to a radio power of just @xmath35whz@xmath11 at the redshift of the cluster . comparing this limit to those achieved by radio surveys of local clusters we note that gavazzi & boselli ( 1999 ) detect galaxies as faint as @xmath36whz@xmath11 in their analysis of the radio luminosity function in the virgo cluster ( a factor of roughly @xmath37 closer to us than cl0939 + 4713 ) using the nvss 1.4ghz survey ( condon et al .
1998 ) .
table 1 properties of the radio galaxy sample cccccl & r.a . &
dec . & @xmath38 & @xmath39 & comments i d & 2 ( j2000 ) & ( @xmath2jy ) & ( @xmath40whz@xmath11 ) & 36 & 09 42 58.16 & + 46 59 53.4 & @xmath41 & 5.9 & 122 & 09 42 57.62 & + 46 59 45.3 & @xmath42 & 15.2 & resolved?224 & 09 42 55.80 & + 46 59 40.4 & @xmath43 & 4.0 & 230 & 09 43 02.32 & + 46 59 28.7 & @xmath44 & 3.8 & @xmath39 assumes @xmath0296 & 09 43 02.81 & + 46 59 24.2 & @xmath45 & 47.6 & resolved?399 & 09 42 57.84 & + 46 59 13.3 & @xmath46 & 3.1 & 426 & 09 42 56.21 & + 46 59 12.0 & @xmath47 & 2.4 & 536 & 09 43 00.03 & + 46 58 53.4 & @xmath48 & 14.5 & 610 & 09 42 57.34 & + 46 58 50.4 & @xmath49 & 11.7 & 640 & 09 42 55.08 & + 46 58 41.7 & @xmath50 & 2.6 & these observations were obtained as part of the early release observations during the science verification phase after the installation of the _ wfpc2 _ on board _ hst_. the observations comprise a total of 21.0ks of integration through the f702w filter ( dressler et al .
the exposures were dithered by integer pixels to allow the removal of defects and cosmic ray events .
the position of the corrective optics within _ wfpc2 _
, to correct for the aberration of the primary mirror , was not optimal during these observations and combined with a higher than ideal operating temperature for the ccds these images are not as cosmetically clean as later exposures with _
nevertheless , they provide deep ( @xmath51 ) imaging of the core of the cluster at 0.10.2@xmath25 resolution in the restframe @xmath52-band .
the reduction and analysis of the f702w exposure of this field is described in more detail in smail et al .
( 1997 ) , who provide a complete catalog of objects detected in the field with photometry calibrated to the @xmath53 passband described by holtzmann et al .
smail et al .
( 1997 ) also discuss visually classified morphologies on the revised hubble scheme for the brighter galaxies within this field down to @xmath54 .
there are a total of 123 galaxies with morphologies from s97 within the joint _
nicmos_/_wfpc2 _ field .
all ten of the radio sources selected from the vla map are included in the s97 catalog and we list the hubble type for each galaxy in table 2 .
we also give the disturbance index ( @xmath55 ) , which is a measure of how disturbed the galaxy seemed compared to the typical appearence of a galaxy with its hubble type locally , where @xmath56 is `` normal '' and @xmath57 is highly disturbed ( s97 ) .
in addition to these objective measures , a subjective interpretation of the source of the disturbance was also given by s97 and is listed in the table ( e.g. chaotic , tidal interaction , etc ) .
we note in particular that a galaxy would be classed as `` merger '' if two or more close nuclei were seen in a common envelope .
these visual estimates were shown to correlate well with the machine - based asymmetry measurements ( s97 ) and have been used to attempt to isolate and study the role of dynamical interactions in triggering the butcher - oemler effect ( e.g. d99 ) .
fig . 2 . the @xmath58@xmath27 color - magnitude distribution for the sources lying in the joint _ wfpc2_/_nicmos _ field .
the @xmath58 color is measured within an aperture of diameter 2.5 times the fwhm of the galaxy in the f160w image .
the large filled points denote those galaxies detected in the vla map , the circled points are other galaxies with spectroscopic identifications in d99 and the small points are the remaining sources selected from the f160w catalog .
note the sequence of red galaxies associated with early - type cluster members which extends down to @xmath59 . to investigate the optical - infrared colors of the galaxies in our sample we have aligned and resampled the _ wfpc2 _ image to the reference frame of the _ nicmos _ exposure .
a comparison of stars and compact galaxies in the resampled f702w exposure and the f160w image shows that they have very similar profiles in the two bands indicating that the resolution is dictated by the pixel sampling of the images .
this suggests that psf variations should not strongly influence our analysis . to confirm this
we have convolved the _ wfpc2 _ and _ nicmos _ images with a model psf for the other instrument generated using tinytim ( krist & hook 1999 ) .
the colors determined from these convolved images show no systematic differences from those measured from the unsmoothed images and hence we use the latter in the following analysis .
to measure colors from our f702w and f160w exposures we have used the iraf task phot and apertures matched to the near - infrared extent of the galaxies .
the diameter of the photometry apertures were taken as 2.5 times the fwhm of each source from the f160w catalog ( @xmath60 in table 2 ) .
the color - magnitude diagram for the field is shown in fig . 2 , with the radio sources and spectroscopically identified galaxies marked .
we give total @xmath27 magnitudes and aperture @xmath58 colors for the ten radio sources in table 2 .
colors of the various @xmath61 lensed galaxies identified by trager et al .
these galaxies are all relatively blue compared to the general field ( fig . 2 ) , @xmath62 for dg#334 at @xmath63 , @xmath64 for the @xmath65 source seen as p1/p2/p3 and @xmath66 for the @xmath67 arclet a0 , although this may suffer contamination from the nearby bright cluster galaxy ( # 426 , fig .
a search of the _ nicmos _ and _ wfpc2 _ images to identify galaxies with extreme colors turned up a very red disk galaxy near the center of the field : # 333 ( fig .
1 ) , this is the reddest , bright object in the field with @xmath68 and @xmath69 ( c.f .
treu et al.1999 ) . ]
we show the f160w and f702w images of this sample in fig . 3 . to study the color variations within the galaxies we have constructed f702w@xmath70f160w color images for each galaxy in our spectroscopic sample .
we show the color images for the relevant radio - detected galaxies in fig .
3 to illustrate the internal color variations within these galaxies .
a comparison sample of confirmed cluster members which are not detected in the radio map is shown in fig .
we note that psf - convolved versions of these images are qualitatively similar , in particular the blue cores at the centers of # 296 , # 399 and # 426 do not appear to result from differences in the psf between the two passbands .
table 2 optical / near - infrared properties of the radio sample cccccccccl & @xmath71 & @xmath27 & ( @xmath72 ) & @xmath60 & morph .
& spectral & ew(h@xmath5 ) & @xmath55 & comments i d & & & & ( @xmath25 ) & type & type & ( ) & & 36 & 0.4119 & 18.20 & @xmath73 & 5.6 & sd & e(a ) & 10.2 & 2 & ew([oii])@xmath74 ; merger 122 & 0.4125 & 17.96 & @xmath75 & 2.7 & sd & a+k & 11.6 & 3 & merger , h@xmath5 measurement uncertain .
224 & 0.4076 & 15.85 & @xmath76 & 5.8 & scd ? & k+a & 6.1 & 2 & merger 230 & ... & 18.80 & @xmath77 & 2.2 & irr & ... & ... & 2 & no spectrum in d99296 & 0.4014 & 16.87 & @xmath78 & 2.5 & e & k+a & 4.6 & 0 & 399 & 0.4109 & 16.14 & @xmath79 & 2.5 & e & k & 0.0 & 0 & 426 & 0.4037 & 16.18 & @xmath80 & 3.9 & e & k & 0.0 & 0 & 536 & 0.4111 & 17.49 & @xmath81 & 4.0 & sc & a+k & 8.6 & 3 & tidal ? 610 & 0.4007 & 15.84 & @xmath79 & 3.4 & sa / s0 & k+a & 4.2 & 1 & merger ?
640 & 0.3324 & 16.27 & @xmath82 & 8.3 & sab & k : & 0.0 & 1 & field galaxy , low s / n spectrum possible e(a ) ?
spectroscopic observations of galaxies within our field are available from the morphs spectral catalog ( d99 ) .
the targets for these observations were selected from the f702w exposure and d99 amassed spectra for 28 galaxies within the _ nicmos _ field , comprising 22 cluster members and 6 field galaxies .
these spectra typically cover the spectral region from 30006000 in the restframe , including any [ oii]@xmath63727 emission but in most cases not h@xmath8 .
the galaxies have all been classified using the spectral classification scheme of d99 as follows : passive galaxies ( k ) ; post - starburst ( k+a / a+k ) , these show no [ oii ] and strong balmer absorption indicative of a cessation of star formation within the last gyr or so ; actively star forming ( e(c)/e(b)/e(a ) ) , e(c ) have spectral properties similar to local spiral galaxies , e(b ) have stronger [ oii ] and are interpreted as starburst galaxies , e(a ) have moderate [ oii ] but stronger balmer lines than e(c ) galaxies . the e(a ) class has been interpreted as dusty starbursts ( p99 ) where the [ oii ] is suppressed by dust within the galaxy ( supported by the observations of anomalously high h@xmath8/[oii ] ratios for these galaxies , see also 1 ) .
the distribution of cluster / field galaxies in the different spectroscopic classes is : k ( 7/2 ) ; k+a ( 6/0 ) ; a+k ( 2/0 ) ; e(a ) ( 5/1 ) ; e(b ) ( 1/0 ) ; e(c ) ( 2/1 ) .
the spectrum of one field galaxy has too low signal - to - noise to allow a reliable classification .
for the ten radio sources in the field , spectra and spectral classifications are available from d99 for nine galaxies ( galaxy # 230 lacks a spectroscopic i d ) and we list these in table 2 .
of these nine galaxies , eight are cluster members and one is a field galaxy ( the sab galaxy # 640 ) .
with a small sample of galaxies our conclusions are by necessity fairly qualitative in nature and this paper should be seen as the prelude to the analysis of a larger sample of radio - selected galaxies in distant clusters by morrison et al .
nevertheless , we will use our observations , and particularly the high resolution near - infrared imaging with _ nicmos _
, to illustrate the general characteristics of radio galaxies in distant clusters and to emphasise the effects of dust in galaxies on their perceived properties .
we have detected ten galaxies above the 3@xmath1 flux limit ( @xmath83jy ) of the vla map which lie within the joint _
nicmos_/_wfpc2 _ field .
the spectroscopic information available from d99 allows us to identify eight of these galaxies as cluster members ( or 35% of the 22 cluster galaxies known within this field ) and a further source as a foreground field galaxy ( 12% of the spectroscopically - confirmed field population within our image ) .
our field size corresponds to a region @xmath84kpc in extent at the cluster redshift and the equivalent radio power limit is @xmath35whz@xmath11
. a comparably deep survey of a similar region in the core of a rich local cluster would typically detect @xmath85 galaxy ( e.g. fig .
5 ) , compared to the eight detected here .
this suggests a substantial increase in the rate of occurence of radio sources in high density environments at moderate redshifts ( morrison et al .
1999 ) .
the galaxies included in the radio - selected sample of cluster members are typically the brighter galaxies in our field ( fig .
2 ) , including the brightest four galaxies ( in the restframe @xmath52-band ) . the whole sample spans a range of magnitudes @xmath8620.8 ( @xmath87@xmath88 ) .
the colors of the fainter radio - detected galaxies are slightly bluer than equivalently luminous , undetected cluster members , with the brighter radio - emitting galaxies having comparable ( red ) colors to the undetected population ( fig . 2 ) .
the fact that we only detect the optically brightest galaxies in the radio map may suggest that the majority of cluster galaxies would be detected if we could increase the sensitivity of the radio observations by a factor of a few ( see fig .
5).kms@xmath11 for the eight radio - detected members compared to @xmath89kms@xmath11 for the 12 undetected members lying within the same field . ]
in our analysis we will assume that the local relationship between star formation rate and radio emission ( e.g. condon 1992 ) holds for spiral galaxies detected in distant clusters .
the validity of the local relationship at least for the distant field appears to be supported by recent observations ( georgakakis et al .
. however , we caution that the exact star formation rates derived from this analysis may have substantial systematic uncertainties .
in particular the claim that the radio to far - infrared correlation varies locally between galaxies in rich clusters and those in the field ( andersen & owen 1995 ) may indicate an environmental influence in the conversion from radio luminosity to star formation rate ( sfr ) resulting from the compression of the galactic magnetic field by the galaxy s motion through the intracluster medium ( icm ) . without a fuller understanding of the state of the icm in distant clusters it is difficult to quantify the extent of this bias on the radio - derived sfrs of distant cluster galaxies .
we have chosen therefore to quote the sfrs determined using the local field relationship and the correction which should be applied if the radio population in distant clusters shows the same behaviour as is claimed for galaxies in rich clusters locally . turning to the restframe @xmath52-band morphologies of the radio sources ( table 2 & fig .
3 ) we see that they fall roughly equally into two broad classes : optically - luminous early - type galaxies ( e and s0/sa ) and mid- to late - type spiral galaxies ( sc sd / irr ) . in their analysis of galaxies in the virgo cluster detected in the nvss radio survey ,
gavazzi & boselli ( 1999 ) found that the most powerful radio sources in spirals were also in mid - type , sb
sc , systems . to compare the optical and radio properties of the galaxies
we plot the 1.4-ghz restframe power versus the optical @xmath52-band luminosity for all spectroscopically - confirmed members in fig . 5 , separated into early- and late - type .
the early - type galaxies show a wider range in their @xmath90 values than the spirals , but span a similar region of the @xmath39@xmath91 plane as radio - selected e / s0 galaxies in the virgo cluster ( fig . 5 ) .
this is as expected if the radio emission from the early - type galaxies is coming from a central agn .
in contrast the radio emission from the spirals is ( by analogy to local systems ) most likely synchotron emission associated with active star formation and leading to a correlation between the radio power and the optical luminosity of the galaxy .
we note that the population of radio - emitting spiral galaxies identified by gavazzi & boselli ( 1999 ) in virgo have radio powers @xmath92 lower at a fixed @xmath52-band luminosity than the galaxies we are detecting in cl0939 + 4713 ( fig . 5 ) .
fig . 5 .
the @xmath39@xmath93 distribution for those galaxies detected in the vla map of cl0939 + 4713 compared to the equivalent distribution for the virgo cluster ( gavazzi & boselli 1999 ) . the top panel compares the distributions for the morphologically - classified spiral and irregular galaxies in both clusters , while the lower panel shows the equivalent comparison for elliptical and s0 members .
we adopt an apparent magnitude of @xmath94 for an @xmath95 ( @xmath96 ) galaxy in the @xmath53-band observed at @xmath0 ( assuming an sed typical of a present - day sbc galaxy ) .
the dotted lines show the radio flux limit of our sample in cl0939 + 4713 .
we indicate the spectroscopically - confirmed member galaxies which are undetected in the vla map as upper limits .
the association of around 60% of the bright radio - selected sample with spiral galaxies suggests that the evolution in the radio populations within clusters runs in parallel to the increasing fraction of blue star - forming galaxies found in these regions at @xmath97 , the butcher - oemler effect .
the rise in the fraction of bright spiral galaxies in clusters such as cl0939 + 4713 is a factor of @xmath20 out to @xmath15 ( dressler et al .
1997 ) . even without the increased radio luminosity of the brightest cluster spiral galaxies ,
this increase in the proportion of spirals is enough to explain the evolution in the number of radio sources seen in the distant clusters ( a factor of @xmath986 , morrison et al .
the increase in the spiral fraction in the clusters is believed to result from infall from the surrounding field associated with the growth of the clusters .
a rise in the radio activity of moderate redshift field galaxies has also been observed and attributed to star formation in these galaxies ( georgakakis et al .
1999 ) . however , to fully understand the evolution of the radio population in the cluster we also need to identify the cause of the increased radio luminosities of the bright spiral galaxies .
we note that our use of a local cluster sample in the comparison should minimise the possible environmental effects which could explain this apparent brightening , and moreover the increase in radio luminosities is substantially above that claimed to result from environmental effects within local clusters ( 3.1 ) .
we now discuss the internal structure of the cluster galaxies in our _ hst _ images ( fig . 3 and 4 ) and the important insights which our _ nicmos _ near - infrared imaging has provided into the nature of these systems . starting with the early - type galaxies we see relatively uniform optical - near - infrared colors ( as shown by their lack of structure in the f702w@xmath70f160w images in fig . 3 and 4 )
a few contain blue cores , which may be associated with central agn , although the details of these features are sensitive to the exact form of the psf in the f702w and f160w images and so we do not discuss them further here . a comparison of the f702w and f160w morphologies for the late - type radio sources in fig .
3 shows that in these systems there are substantial differences in the appearence of these galaxies between the restframe @xmath52- and @xmath28-bands .
the near - infrared morphologies are considerably more relaxed and symmetrical .
however , the lack of a morphological classification scheme for local galaxies based on near - infrared imaging means that is not useful to attempt to reclassify the galaxies on the basis of the f160w images . more interestingly ,
the f702w@xmath70f160w images in fig . 3 show that in a large number of cases the irregular structures seen in the rest - frame @xmath52-band within the central regions of the late - type galaxies are replaced by a much more regular appearence in the near - infrared and so must be due to very red structures ( visible as white regions in fig . 3 ) .
these features are substantially redder than the outskirts of the galaxies ( up to @xmath991.5 ) and could represent differences in the stellar populations within the galaxies or simply dust obscuration . to distinguish between these alternatives we concentrate on the large fraction of galaxies whose integrated spectra show post - starburst stellar populations .
the lack of emission lines in the optical spectra indicates that there is no visible star - formation in the outskirts of these galaxies , which also show the red optical - infrared colors typical of an evolved stellar population ( see the integrated colors in fig . 2 ) .
in this situation it is very difficult to make the stellar population in the central regions of the galaxies substantially redder than the outskirts without invoking a contrived scenario .
in contrast , dust obscuration can easily produce both the apparently red colors and its structured distribution .
we propose therefore that large quantities of dust are probably a common feature of these galaxies .
the small scale features seen in the optical images of these galaxies which we claim are due to dust obscuration have been previously interpreted as arising from dynamical disturbance of these galaxies , suggesting a dynamical origin for their star formation activity ( s97 ) . to understand the consequences of this change in interpretation of these galaxies we focus on those galaxies classified in the optical as showing signs of morphological disturbance by s97 .
there are a total of eight spectroscopically confirmed members in our sample with disturbance indices of @xmath100 indicating abnormally disturbed morphologies , all of these are disk galaxies and they are typically classified as mergers or tidally disturbed , half of them are detected in our radio map .
these galaxies exhibit substantially more regular morphologies in the near - infrared ( figs . 3 & 4 ) ,
in particular the apparently double structures seen in the f702w morphologies of the radio sources # 36 , # 122 and # 536 all disappear in the f160w passband .
we also see signs of dust obscuration in the disturbed galaxies which are not radio sources ( # 171 , # 369 , # 497 and # 622 ) , although arguably not as extreme as the radio - selected sample . in a few cases
there are other signatures of interactions in the f702w images apart from double nuclei , etc , such as low - surface brightness tidal features or companion galaxies ( e.g. # 224 or # 497 ) , which indicate that these are true merging or interacting systems .
but , in the majority of the galaxies their high disturbance index was based upon the apparent presence of multiple nuclei in the central regions of the galaxies ( see 1 ) .
we suggest that in fact dust obscuration is the correct explanation for their irregular appearence in the restframe @xmath52-band which results in their classification as disturbed or merging .
we estimate that of the eight confirmed members with @xmath100 based on the f702w imaging , at least half ( # 36 , # 122 , # 171 , # 622 ) would have @xmath101 if classified from the f160w images , suggesting that the occurence of highly disturbed galaxies in the core of cl0939 + 4713 may have been overestimated by up to a factor of two .
we conclude from our comparison of the near - infrared and optical morphologies of a small sample of confirmed cluster members that the frequency of galaxies showing small - scale signatures of disturbance or interaction is probably substantially less than has been suggested from optical studies ( e.g. s97 ) due to obscuration by dust .
this effect acts in addition to the standard luminosity bias in optically - selected samples towards actively star - forming galaxies ( the usual justification given for undertaking near - infrared surveys of distant clusters , e.g. barger et al .
the expectation of previous optically - based studies was that dust obscuration would not be a significant bias when comparing the distant morphological samples with more local examples as they were comparing effective restframe @xmath52-band observations in both cases ( s97 ) .
however , this expectation seems to have been misleading due to the wider prevelance of dust in cluster galaxies at @xmath102 , perhaps associated with the increased activity in galaxies at these epochs .
looking now at the spectral properties of the radio - selected cluster members we see something striking : five of the galaxies have spectral features which class them as post - starburst ( a+k / k+a ) systems using the scheme of p99 ( table 2 ) .
moreover , these five radio sources show the strongest balmer absorption lines of the eight post - starburst galaxies in our cluster sample ( they are also the brightest five post - starburst galaxies ) .
there are no post - starburst field galaxies in our sample .
the detection of post - starburst galaxies in the radio waveband is surprising .
these galaxies are selected to have no detectable [ oii ] emission and hence are expected to have no on - going massive star formation ( observations of the spectral region including any h@xmath8 emission in these galaxies would make this statement more robust ) .
in contrast , the radio emission is most easily explained as arising from massive star formation ( although some contribution from an obscured radio - loud agn can not be ruled out with our present data ) .
the expected lifetime of the radio emission after the starburst is @xmath103gyr , thus if they are post - starburst systems these galaxies must all be seen _ just _ after the cessation of their star formation .
this seems very unlikely as the length of time in which the post - starburst signature ( strong h@xmath5 absorption ) is visible is @xmath104gyr and we would thus expect to see an order of magnitude more radio - quiet k+a galaxies , where the radio emission had decayed but the balmer absorption is still visible . this population is not observed .
therefore , we conclude that massive stars are currently being formed in these galaxies . where is the site of the active star formation in the a+k / k+a cluster galaxies ?
the optical spectra typically sample the whole of the bulge and disk of the galaxy ( see d99 ) and thus it is unlikely that we are missing emission from the outskirts of the galaxy .
we suggest instead that the star formation is occuring in the central regions of these galaxies but is hidden from view by the dust which we see there in our f702w@xmath70f160w images .
this possibility was raised in p99 who stated that the most extreme e(a ) galaxies might appear as post - starburst systems ( a+k / k+a ) due to obscuration of the emission regions .
the reddened regions lie in the central 510kpc of the galaxies ( fig . 3 & 4 ) and have optical near - infrared colors which are typically @xmath105 magnitude redder than the outskirts of the galaxies .
this would indicate optical extinctions of at least @xmath106mag to the population of disk stars ( let alone the active sites of star formation ) although this estimate is highly uncertain due to resolution effects .
four of the five radio - selected post - starburst galaxies are disk galaxies and we can estimate their star formation rates ( sfr ) from their radio power using the calibration of condon ( 1992 ) . we find that these galaxies are typically forming massive stars ( @xmath107 ) at rates of @xmath108@xmath109 yr@xmath11 , with the highest sfrs being 30@xmath109 yr@xmath11 for # 122 and # 536 .
this assumes no boosting of the radio emission due to compression of the galactic magnetic fields by the cluster icm , which could amount to a factor of @xmath110 ( andersen & owen 1995 ) . combining these sfrs with the @xmath52-band luminosities of the galaxies we can predict their [ oii ] fluxes if the star formation regions are unobscured using the relations given in kennicutt ( 1992 ) .
we can then roughly estimate the amount of extinction needed for this line to be undetected in the optical spectra , ew([oii])@xmath111 .
we find that the star forming regions would need to suffer @xmath112mag of extinction for the [ oii ] line to be undetected in the d99 spectra ( @xmath113mag if we assume the radio fluxes are artificially raised ) .
this amount of obscuration is not unreasonable given the extent of the dust reddening seen in the f702@xmath70f160w images ( figs . 3 & 4 ) .
we note that if we assume this dust is cold , @xmath114k , then the predicted sub - mm fluxes of these galaxies will be only @xmath115mjy at 850@xmath2 m .
this is well below the confusion limit of current sub - mm instrumentation in blank fields .
there are also examples of proposed dusty galaxies within our sample which are not detected in the radio map .
in particular three of the cluster galaxies have e(a ) spectral features and evidence for dust in their f702@xmath70f160w images ( fig . 4 , # 177 , # 369 and # 622 ) .
these galaxies are optically fainter than the e(a ) galaxy which is detected in the radio , # 36 , and it is possible that their lower overall luminosities mean that their radio emission is below the limit of the vla map . interestingly , although the e(a ) spectral class was interpreted as dusty starbursts in p99 , the radio map places upper limits on the current rate of massive star formation in the undetected e(a ) galaxies of @xmath116 yr@xmath11 , with similar limits applying to the galaxies with e(c ) spectra in the field ( e.g. # 165 and # 497 in fig . 4 ) .
taking the possible boosting of the radio emission due to compression of the galactic magnetic fields by the cluster icm ( 3.1 ) into account these sfr limits could be dropped by a factor of two or more .
we see a combination of post - starburst spectral features , radio emission and apparently large quantities of dust within the spiral galaxies in our sample .
these characteristics can be most easily accomodated if they exist in spatially distinct regions of the galaxies , as suggested by our color maps .
the radio emission would arise from massive star formation in the dust obscured center of the galaxy , while the a stars trace the remanents of an earlier ( @xmath85gyr ) strong star formation episode within the galactic disk .
there is weak evidence in our sample for a correlation between h@xmath5 line strength and radio luminosity which would suggest that there is a causal link between the numbers of a stars and the current ( obscured ) star formation rate , probably as a result of a single triggering mechanism .
this triggering mechanism could be a cluster - related phenomenum , a view which is supported by the lack of obvious post - starburst systems in samples of faint radio sources in the field ( hammer et al .
1995 ; l.l .
cowie , priv . comm . ) .
thus , we may be seeing a more dynamic situation including cluster - driven processes , whereby interactions between the intracluster medium and the cold molecular gas within the galactic disks accelerates the collapse and evaporation of the star forming regions .
the effects of the icm on the galactic disk are expected to be a function of the local density in the disk with lower density regions in the outskirts of the disks being most severely effected , while star formation continues almost uneffected in the inner regions of the galaxy .
thus environmental processes could result in a combination of post - starburst and starburst within individual cluster galaxies .
studies of the spatial distribution and kinematics of larger samples of radio - selected galaxies may be able to discriminate between the different physical processes which could be operating ( morrison et al .
1999 ) .
alternatively , perhaps this activity may be triggered by a more general physical process , unrelated to the cluster environment .
indeed , similar behaviour is seen in the models of interacting field galaxies published by mihos & hernquist ( 1996 ) ( although we have argued that a substantial fraction of apparently disturbed cluster spiral galaxies probably arise from dust obscuration , rather than true interactions ) . for a range of models for tidal interactions between galaxies mihos & hernquist
predict that spatially - extended enhanced star formation will first occur in the disk of the galaxy .
this is followed on a timescale of a @xmath1170.3gyr by funnelling of any remaining gas into the center of the galaxy where it reaches very high densities and triggers an ( obscured ? ) nuclear starburst .
again , more detailed investigation of the kinematics and internal dynamics of these galaxies may provide useful diagnostics to distinguish between the competing models .
the picture which is appearing is one in which galaxies which are visibly forming stars appear to have modest star formation rates , @xmath118 yr@xmath11 ( @xmath107 ) , while more vigorous star formation , @xmath119@xmath120 yr@xmath11 is on - going in highly obscured regions in the centers of what appear to be post - starburst galaxies . if the radio fluxes of these distant galaxies are increased by the same environmental effects which are claimed to boost the radio emission of local cluster galaxies and to the same extent ( andersen & owen 1995 ) then the quoted sfrs should be reduced by a factor of @xmath121 .
sensitive observations to search for h@xmath8 emission ( or emission lines further into the near - infrared ) will provide a stronger test of the star formation rates in these galaxies , the extent of obscuration within them and the degree to which their radio emission deviates from that seen in the distant field population due to environmental processes .
finally we note that by having post - starburst and starburst phases occuring in parallel within a single galaxy it is possible to reduce the high fraction of galaxies ( @xmath122% , barger et al .
1996 ) which must pass through this phase to explain the large numbers of post - starburst systems seen within distant clusters ( barger et al . 1996 ; p99 ) .
in summary , we have highlighted the role that dust plays in obscuring our view of the morphologies and star formation properties of distant galaxies .
we have shown that selection at faint radio fluxes provides a powerful technique for identifying spectrally - classified post - starburst and starburst galaxies crucial to our understanding of the evolutionary cycle of galaxies in clusters ( e.g. p99 ) . using this new tool we can undertake statistical studies to search for the mechanism responsible for triggering and/or quenching of this activity ( p99 ; morrison et al . 1999 ) .
the main conclusions of our study are : @xmath123 we present optical and near - infrared imaging with _
hst _ of a region in the core of the distant rich cluster cl0939 + 4713 at @xmath0 .
@xmath123 we compare the high resolution near - infrared and optical morphologies of a sample of 22 cluster members .
this shows that while the broad morphological classifications are similar , the frequency of galaxies showing small - scale signatures of disturbance or interaction is substantially less when using the near - infrared imaging .
we suggest that obscuration by dust has led optical studies to conclude that an anomalously high fraction of galaxies have suffered small scale disturbances .
this may point to a wider prevelance of dust in cluster galaxies at @xmath102 , perhaps as a result of the increased star formation activity at these epochs .
@xmath123 using a very deep vla 1.4-ghz map of this area we select a sample of ten radio galaxies in our field above a flux limit of @xmath124jy , equivalent to a restframe 1.4-ghz power of @xmath125 whz at the cluster redshift .
the faint radio sources are associated with two populations of cluster galaxies : luminous early - type galaxies , the radio emission from these galaxies originates from agn ; and dusty galaxies with late - type spiral morphologies , the radio emission here is expected to trace massive star formation ( although the exact relationship between the radio luminosity and the sfr is uncertain due to environmental influences ) .
@xmath123 the optical spectral characteristics of the late - type radio members are highly unusual they appear to be post - starburst systems .
we suggest that these galaxies are not post - starburst but in fact host highly obscured star - formation and starbursts .
this is supported by our high resolution optical near - infrared imaging which shows that the central regions of these galaxies are heavily dust enshrouded .
if they continued for several @xmath126myrs , these dust enshrouded nuclear - starbursts could build a substantial stellar bulge in the remanent of these luminous disk galaxies , possibly contributing to the formation of the s0 population in the clusters ( mihos & hernquist 1996 ) .
@xmath123 we emphasise that a lack of consideration of the effects of dust on the morphological and spectral properties of galaxies , even at the relatively modest redshifts discussed here , may lead to an incomplete understanding of the cycle of star formation in cluster galaxies
. further work on this subject is urgently required .
@xmath123 this paper is the prelude to a larger study of the _ hst _ morphologies and spectral properties of radio - selected galaxies in cl0939 + 4713 and other high redshift clusters ( morrison et al .
we expect that extensive studies with higher spatial resolution and wider wavelength coverage in the near future using the upgraded vla and the next generation of millimeter interferometers will provide new insights into the prevelance and distribution of dust in distant galaxies .
this paper is based upon observations obtained with the nasa / esa _ hubble space telescope _ which is operated by stsci for the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 and on observations from the very large array which is operated by associated universities , inc .
, under a cooperative agreement with the national science foundation .
we thank lisa storrie - lombardi and lin yan for help with _ nicmos _ photometry and giuseppe gavazzi and alessandro boselli for kindly providing the data on radio galaxies in virgo in electronic form .
we also thank john hill and the _ nicmos _ gto team for allowing us to observe this target .
we acknowledge the use of the pedestal estimation and quadrant equalization software developed by roeland p. van der marel .
we thank an anonymous referee for detailed comments on this work and amy barger , andrew blain , warrick couch , len cowie , alan dressler , alastair edge , gus oemler and especially bianca poggianti for useful conversations .
irs acknowledges support from the royal society and rji from pparc .
fig . 1 .
the vla 1.4-ghz map overlayed as a contour plot on the _ hst _ _ nicmos _ f160w mosaic .
the radio sources are identified in large font using the numbering scheme of smail et al .
the galaxies labelled with the smaller font represent other spectroscopically identified galaxies ( with the exception of # 333 ) from the catalog of dressler et al .
( 1999 ) , these are similarly numbered using the s97 identifications .
the dashed line shows the overlap on this field of the _ wfpc2 _
f702w exposure from s97 .
the radio contours are at 12 , 25 , 50 , 100 and 200@xmath2jy per beam .
north is top and east is left in this figure .
the groups of _ nicmos _ f160w , f702w@xmath70f160w and _ wfpc2 _ f702w images for each of the ten galaxies selected from the vla 1.4-ghz map of cl0939 + 4713 . the f160w and f702w images provide restframe @xmath28- and @xmath52-band views of the galaxies and these are shown with a linear intensity scale , while the color image is constructed from the difference of the @xmath128-scaled f160w and f702w images . in the f702w@xmath70f160w color image lighter regions represent redder areas of the galaxy , the typical color range is roughly @xmath1293.2 ( black to white ) .
the panels are labelled with the catalog number , morphological type , spectral type and disturbance index ( from s97 and d99 ) .
each panel is @xmath130 ( equivalent to 65kpc square at the cluster redshift ) with north top and east left and the panels are ordered on morphology .
note that # 640 lies in the extreme corner of the _ wfpc2 _ field and # 230 falls on a chip - boundary in the f702w exposure .
all of the galaxies are confirmed cluster members with the exception of # 640 which is a foreground field galaxy and # 230 which has no spectroscopic identification .
4 . a comparison sample of spectroscopically - confirmed cluster galaxies which show no detectable radio emission in the vla map above a 1.4-ghz flux of @xmath124jy .
this figure includes all the spectrally - active members of this sample and again is ordered on morphology .
the panels are labelled with the galaxy i d , morphology and disturbance index , d , from s97 and the spectral classification from d99 . | we present deep _ hubble space telescope _ _ nicmos _ near - infrared and _ wfpc2 _ optical imaging of a small region in the core of the distant rich cluster cl0939 + 4713 ( @xmath0 ) .
we compare the optical and near - infrared morphologies of cluster members and find apparent small - scale optical structures within the galaxies which are absent in the near - infrared .
we conclude that strong dust obscuration is a common feature in the late - type galaxies in distant clusters .
we then concentrate on a sample of ten faint radio galaxies lying within our _ nicmos _ field and selected from a very deep 1.4-ghz vla map of the cluster with a 1@xmath1 flux limit of 9@xmath2jy . using published data we focus on the spectral properties of the eight radio - selected cluster members and
show that these comprise a large fraction of the post - starburst population in the cluster .
the simplest interpretation of the radio emission from these galaxies is that they are currently forming massive stars , contradicting their classification as post - starburst systems based on the optical spectra .
we suggest that this star formation is hidden from view in the optical by the same obscuring dust which is apparent in our comparison on the optical and near - infrared morphologies of these galaxies .
we caution that even in the restframe optical the effects of dust can not be ignored when comparing samples of distant galaxies to low - redshift systems , particularly if dust is as prevelant in distant galaxies as appears to be the case in our study . |
deep surveys at fir and submilimeter wavelengths have been carried out in order to investigate the nature of dust - enshrouded galaxies at high redshift . as a contribution to this field ,
our group made a deep fir survey using the isophot camera on board the _ infrared space observatory _
( iso ) satellite ( kawara et al . 1998 ; matsuhara et al . 2000 ) .
mapping at 90@xmath0 m and 170@xmath0 m of two @xmath1 fields in the lockman hole ( lh_ex and lh_nw ) , a region exhibiting the lowest hi column density in the sky ( lockman et al .
1986 ) , resulted in the detection of 36 sources with @xmath2150mjy and 45 sources with @xmath3150mjy .
given the relatively large size of the isophot beam at 170@xmath0 m ( @xmath490@xmath5 ) , we have obtained opt / nir images and spectra using telescopes on mauna kea and 6 cm radio continuum maps using the vla ( yun et al .
2002 ) to identify the most likely source of the 170@xmath0 m emission . here
we report our initial identifications of the brightest of the isophot 170@xmath0 m sources .
redshifts of 35 fir source candidates were determined using optical spectra obtained with esi on keckii during three observing runs in 2000 march and 2001 january .
infrared luminosities , @xmath6 , were then estimated by using the isophot fluxes and assuming an sed similar to that of arp220 .
we found one hyperluminous infrared galaxy ( hylig : @xmath7 ) at @xmath8 , 11 ultraluminous infrared galaxies ( uligs : @xmath9 ) at @xmath10 , 12 luminous infrared galaxies ( ligs : @xmath11 ) , and 11 galaxies with @xmath12 . except for one lig at @xmath13 ,
all of the galaxies with @xmath14 ) are at @xmath15 .
the mean redshift for all sources is @xmath16 .
the low - resolution esi spectra were used to determine the optical spectral - type of the candidate isophot sources . following procedures used by murayama & taniguchi ( 1998 ) ,
the spectra were classified into four types agns , liners , hii - type , and early - type ( without emission lines ) .
the hylig at @xmath8 was found to be a quasar .
one ulig had an early - type spectrum and 10 uligs are hii galaxies . among the remaining 23 lower - luminosity sources ,
there was one early - type galaxy , one seyfert 2 , 10 liners and 11 hii galaxies .
thus , based on our low - resolution esi optical spectra most of the isophot 175@xmath0 m sources appear to be powered primarily by star formation , consistent with the conclusion reached from an analysis of isophot number counts by matsuhara et al .
( 2000 ) that most of the isophot sources are star - forming galaxies at @xmath17 . | we report on optical and near - infrared ( nir ) follow - up spectroscopy of faint far - infrared ( fir ) sources found in our deep fir survey by kawara et al .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
inverse heat source problems appear in many branches of engineering and science .
a typical application is for instance an accurate estimation of a pollutant source , which is a crucial environmental safeguard in cities with dense populations ( e.g. for instance @xcite and @xcite ) .
these inverse problems are severely ill - posed , involving a strongly time - irreversible parabolic dynamics .
their mathematical analysis is difficult and still a widely open subject . in the present work ,
we are concerned with sources located at the initial time . in an earlier paper gilliam and
martin @xcite considered the problem of recovering the initial data of the heat equation when the output is measured at points discrete in time and space .
they observed that this problem is linked to the theory of dirichlet series and a solution is found in the one dimensional case .
in @xcite , gilliam , lund and martin provide a simple and extremely accurate procedure for approximating the initial temperature for the heat equation on the line using a discrete time and spatial sampling .
this procedure in based on `` sinc expansion '' .
later , in @xcite , the same authors give a discrete sampling scheme for the approximate recovery of initial data for one dimensional parabolic initial - boundary value problems on bounded interval .
the problem of recovering the initial states of distributed parameter systems , governed by linear partial differential equations , from finite approximate data , was studied by gilliam , mair and martin in @xcite .
li , osher and tsai considered in @xcite the inverse problem of finding sparse initial data from the sparsely sampled solutions of the heat equation .
they prove that pointwise values of the heat solution at only a few locations are enough in an @xmath0 constrained optimization to find the initial data . in a recent work ,
de vore and zuazua @xcite studied the problem of approximating accurately the initial data if the one - dimensional dirichlet problem from finite measurements made at @xmath1 .
they proved that , for suitable choices of the point location of the sensor , @xmath2 , there is a sequence @xmath3 of time instants that guarantees the approximation with an optimal rate , of the order of @xmath4 in the @xmath5-sense , depending on the sobolev regularity of the datum being recovered .
it is worth mentioning that the orthogonality method in @xcite leads to a best approximation in space , while in @xcite the authors show a best approximation in time .
the determination of the initial distribution in the heat equation on a flat torus of arbitrary dimension was considered by danger , foote and martin @xcite .
they established that the observation of the solution along a geodesic determines uniquely the initial heat distribution if and only if the geodesic is dense in the torus .
their result was obtained by using a fourier decomposition together with results from the theory of almost periodic functions .
similar ideas were independently employed in @xcite in the context of the approximate controllability and unique continuation for the heat equation along oscillating sensor and actuator locations .
let @xmath6 be the solution of the heat equation in the whole space @xmath7 .
nakamura , saitoh and syarif @xcite showed that the initial distribution is determined and simply represented by the observations @xmath8 and @xmath9 , @xmath10 , @xmath11 , for some fixed @xmath12 . to our knowledge
there are only few works dealing with the stability issue . in a series of papers by saitoh and al .
@xcite , based on reznitskaya transform and some properties of bergman - selberg spaces , they obtained lipschitz stability estimate from a point or a boundary observation .
we consider the initial - boundary value problem , abbreviated to ibvp in the sequel , for the one dimensional heat equation @xmath13 it has a unique solution @xmath14 whenever @xmath15 .
this solution can be written in term of fourier series as follows @xmath16 where @xmath17 is the @xmath18-th fourier coefficient of @xmath19 : @xmath20 given a point @xmath21 for the placement of the sensor , we address the question of reconstructing the initial distribution @xmath19 from @xmath22 , @xmath23 . in light of , setting @xmath24 , we see that the actual problem is reduced to one of recovering the sequence @xmath25 from the sum of the corresponding dirichlet series @xmath26 for this to be the case @xmath2 has to be chosen in a strategic way so that @xmath27 for all @xmath28 .
section 2 is devoted to establish stability estimates for the problem of recovering the coefficients @xmath29 of a general dirichlet series @xmath30 from its sum . here
@xmath31 is a strictly increasing sequence of non negative real numbers , converging to infinity .
the behavior of this problem depends on whether the series @xmath32 converges .
* when @xmath32 converges , @xmath33 admits a biorthogonal family and this simplifies the analysis .
* the case @xmath34 is much harder and the stability estimates we obtain are weaker . in both cases ,
we prove stability estimate of logarithmic or hlder type and to carry out our analysis we need to compare @xmath35 with @xmath36 , @xmath37 is given .
in the case @xmath34 , a gap condition on the sequence @xmath31 is also necessary in our analysis ( see in subsection [ sb2.2 ] ) . in section 3 ,
we apply the results obtained in the section 2 to the problem of determining the initial heat distribution in a one dimensional heat equation from an overspecified data .
we get a various stability estimates of hlder or logarithmic type .
our results include fractional heat equations of any order .
we also establish a boundary observability inequality with an output located at one of the end points .
this later enables us to obtain a logarithmic stability estimate for the inverse problem of recovering the initial condition from a boundary measurement . in section 4 ,
we first consider the particular case of the ( fractional ) heat equation in a d - dimensional rectangle @xmath38 .
we show that if the dirichlet eigenvalues of the laplacian in @xmath38 are simple . ] , then , in some cases , the determination of the initial heat distribution can be reduced to the one dimensional case .
this is achieved when the measurements consist in the values of the solution of the heat equation on @xmath39 affine @xmath40-dimensional subspaces .
next , we revisit the case when overdetermined data is an internal or a boundary observation .
we comment the existing results and show that these laters can be improved using recent observability inequalities .
the unit ball of a banach space @xmath41 is denoted by @xmath42 .
@xmath43 , @xmath44 , is the usual banach space of complex - valued sequences @xmath45 such that the series @xmath46 is convergent .
we equip @xmath47 with its natural norm @xmath48 @xmath49 denotes the usual banach space of bounded complex - valued sequences @xmath45 , normed by @xmath50 for @xmath51 , the space @xmath52 is defined as follows @xmath53 where @xmath54 .
@xmath55 is a hilbert space when it is equipped with the norm @xmath56
we limit our study to dirichlet series whose coefficients consist in sequences from @xmath47 , @xmath57 .
we pick a real - valued sequence @xmath31 satisfying @xmath58 and @xmath59 as @xmath18 goes to @xmath60 . to @xmath61 , @xmath57 ,
we associate the dirichlet series @xmath62 let @xmath63 and @xmath64 be the conjugate exponent of @xmath65 . observing that @xmath66 we get by applying hlder s inequality @xmath67 also , @xmath68 therefore , the series @xmath69 converges for @xmath70 , for any @xmath71 , @xmath57 . from the classical theory of dirichlet series
( see for instance @xcite ) , we know that @xmath72 has an analytic extension to the half plane @xmath73 . since a dirichlet series is zero if and only if its coefficients are identically equal to zero , we conclude that the knowledge of @xmath72 in a subset of @xmath73 having an accumulation point determines uniquely @xmath74 . in other words , if @xmath75 has an accumulation point and @xmath72 vanishes on @xmath76 , then @xmath77 .
the most interesting case is when @xmath78 .
we can define in that case the operator @xmath79 by @xmath80 here , @xmath81 is the banach space of bounded continuous function on @xmath82 , equipped with the supremum norm @xmath83 then @xmath79 is an injective linear contractive operator .
similarly to entire series , we address the question to know whether it is possible to reconstruct the coefficients of a dirichlet series from its sum .
this is always possible if the values of the sum is known in the half plane @xmath73 .
more specifically , we have the following formula ( see a proof in @xcite ) : for @xmath84 and @xmath85 , @xmath86 in the present section we aim to establish the modulus of continuity , at the origin , of the inverse of the mapping @xmath87 , where @xmath76 is a subset of @xmath88 . rouhly speaking , we seek an estimate of the form @xmath89 for @xmath74 in some appropriate subset of @xmath47 , @xmath78 or @xmath90 , where @xmath91 is a continuous , non decreasing and non negative real - valued function satisfying @xmath92 .
we discuss separately the cases @xmath93 and @xmath94 .
it is worthwhile recalling that the nature of the series @xmath95 is related to mntz s theorem saying that the closure of the vector space spanned by @xmath96 is dense in @xmath97 if and only if @xmath98 .
usually this theorem is stated in the following equivalent form : @xmath99 is dense in @xmath100)=\{\varphi \in c_0([0,1]);\ ; \varphi ( 0)=0\}$ ] if and only if @xmath101 .
let @xmath31 be the sequence of the dirichlet eingenvalues of the laplacian on a bounded domain of @xmath7 , that we assumed simple . from (
* lemma 3.1 , page 229 ) , @xmath102 . therefore
, in that case @xmath103 holds if and only if @xmath104 .
the following lemma will be useful in the sequel .
henceforth , @xmath105 [ lemma2.1 ] let @xmath106 .
then @xmath107 and @xmath108 converges to @xmath72 in @xmath109 as @xmath110 . from cauchy - schwarz s inequality @xmath111 but @xmath112
hence @xmath107 and @xmath113 similarly @xmath114 thus @xmath115 as @xmath110 , for any @xmath116 $ ] . as @xmath117 , @xmath118 , we apply lebesgue s dominated convergence theorem in order to get that @xmath108 converges to @xmath72 in @xmath109 .
let @xmath119 be the closure in @xmath109 of the vector space spanned by @xmath96 . by ( * ?
* theorem in page 24 ) , @xmath119 is a proper subspace of @xmath109 .
additionally , @xmath120 possesses a biorthogonal set @xmath121 in @xmath109 : @xmath122[2.3 ] we assume that there are constants @xmath123 ,
@xmath124 and @xmath125 such that @xmath126 from in ( * ? ? ?
* formula @xmath127 ) , @xmath128 for some constant @xmath129 that can depend only on @xmath130 . in light of
, we get @xmath131 therefore , by cauchy - schwarz s inequality , @xmath132 , we can pass to the limit , as @xmath110 , in .
we get @xmath133 then , implies @xmath134 , @xmath135 but @xmath136 by .
consequently , @xmath137 let @xmath51 and @xmath138 be two given constants . by inequality
we have , for any @xmath139 , @xmath140 that is , @xmath141 we need the following result to pursue our analysis . it is stated as theorem 5.1 in @xcite .
[ theorem2.1 ] let @xmath142 .
there is a constant @xmath143 depending only on @xmath144 ( and not on @xmath145 , @xmath146 and the length of @xmath65 ) so that @xmath147 for every @xmath148 and every lebesgue - measurable set @xmath149 $ ] of lebesgue measure at least @xmath150 .
[ corollary2.1 ] let @xmath151 be a lebesgue - measurable set of @xmath152 $ ] of positive lebesgue measure .
there is a constant @xmath39 , that can depend on @xmath31 , @xmath151 and @xmath153 , so that , for any @xmath154 , @xmath155 we proceed similarly as in the beginning of the proof of ( * ? ? ?
* corollary 5.2 ) .
let @xmath156 and @xmath157.\ ] ] then @xmath158 we recall that @xmath159 we get by applying theorem [ theorem2.1 ] @xmath160 on the other hand , @xmath108 converges uniformly to @xmath72 in @xmath82 .
this is an immediate consequence of the following estimate @xmath161 therefore , is obtained by passing to the limit , as @xmath110 , in .
now , estimate in yields , where @xmath151 is a given lebesgue - measurable set of @xmath152 $ ] of positive lebesgue measure , @xmath162 let @xmath163 be the greatest integer satisfying @xmath164 such an @xmath163 exists provided that @xmath165 is sufficiently small .
a straightforward computation shows that @xmath166 by taking @xmath167 in , we get that there exist @xmath168 so that @xmath169 when @xmath170 , @xmath171 a combination of and implies @xmath172 we sum up our analysis in the following theorem . [ theorem2.2 ]
we assume that assumption is satisfied .
let @xmath151 a lebesgue - measurable set of @xmath152 $ ] of positive lebesgue measure , @xmath138 and @xmath173 .
there exists a constant @xmath129 , that can depend on @xmath151 , @xmath130 , @xmath174 and @xmath175 , so that , for any @xmath176 , @xmath177 we observe that @xmath178 when @xmath179 .
therefore , @xmath180 if @xmath179 .
[ remark2.1 ] 1 ) in light of and , we have the following lipschitz stability estimate , where @xmath143 is a constant depending only on @xmath130 , @xmath181 here the left hand side of this inequality is seen as a an @xmath182-weighted norm of @xmath74 .
\2 ) starting from , we can prove the following estimate @xmath183 for any @xmath51 .
\3 ) it is possible to establish a hlder stability estimate .
this is can be done by substituting in theorem [ theorem2.2 ] @xmath184 by the following subspace @xmath185 with @xmath186 and @xmath187 .
a proof of a similar result will be detailed in the next subsection .
\4 ) a lipschitz or a hlder stability estimate is not true in general .
indeed , let us assume that we have an estimate of the form , where @xmath188 , @xmath189 let @xmath190 be the usual orthonormal basis of @xmath191 .
that is @xmath192 , where @xmath193 is the kronecker symbol . letting @xmath194
, we get by a straightforward computation @xmath195 since @xmath196 , if is true then we would have from @xmath197 the particular choice of @xmath198 , @xmath199 in yields @xmath200 but this inequality can not be true if @xmath201 .
we pick @xmath202 and we set @xmath203 since @xmath204 , we have @xmath205 the choice of @xmath206 gives @xmath207 more generally , we have @xmath208 and then @xmath209 so an induction argument leads to the following estimate @xmath210 and @xmath211 .
therefore @xmath212 , @xmath213 .
if @xmath214 then implies @xmath215 we introduce the weighted @xmath0 space , where @xmath173 , @xmath216 we equip @xmath217 with its natural norm @xmath218 let @xmath219 . in light of
, we have @xmath220 hence @xmath221 let us assume that the sequence @xmath31 obeys to the following assumptions : there are four constants @xmath222 , @xmath223 , @xmath186 and @xmath224 so that @xmath225 under these assumptions , @xmath226 therefore , yields @xmath227 if @xmath228 is sufficiently small , we denote by @xmath229 the greatest positive integer such that @xmath230 let @xmath231 . since @xmath232 we have @xmath233 by a straightforward computation .
we end up getting @xmath234 for some @xmath235 .
when @xmath236 , we have @xmath237 hereafter we used that @xmath238 , @xmath239 , which a consequence of the linearity of @xmath79 . in light of and , we can state the following result .
[ theorem2.4 ] we assume that assumption fulfills .
let @xmath138 .
there exists a constant @xmath129 , that can depend only on @xmath240 and @xmath174 , so that , for any @xmath241 ,
@xmath242 we are now going to show that , even in the present case ( @xmath98 ) , it is possible to establish a hlder stability estimate .
we pick @xmath243 , @xmath244 an non negative integer and we recall that @xmath108 is given by @xmath245 let @xmath246 , @xmath247 .
we introduce the following vandermonde matrix @xmath248 by setting @xmath249 and @xmath250 , we get in a straightforward manner that @xmath251 or equivalently @xmath252 .
if @xmath253 and @xmath254 , then @xmath255 from the proof of ( * ? ? ?
* theorem 1 ) , we obtain @xmath256 therefore , under assumption , we get after some technical calculations @xmath257 where @xmath258 is the same as in . hence , entails @xmath259 for @xmath125 and @xmath37 , we introduce the following weighted @xmath0-space : @xmath260 we equip this space with its natural norm @xmath261 let @xmath174 be a non negative constant and @xmath262 ( @xmath258 is the same as in ) . assuming that @xmath263 , we obtain in light of @xmath264 therefore , we find an integer @xmath265 so that for any @xmath266 , @xmath267 we derive by minimizing with respect to @xmath244 the following hlder stability estimate .
[ theorem2.5 ] we assume that fulfills .
let @xmath138 , @xmath125 and @xmath262 .
there exist two constants @xmath129 and @xmath85 , that can depend only on @xmath130 , @xmath174 , @xmath268 and @xmath269 , so that , for any @xmath270 , @xmath271
we come back to the one dimensional heat equation . we consider again the ibvp @xmath272 the solution of the ibvp is given by @xmath273 where @xmath17 is the fourier coefficient of @xmath15 : @xmath20 from (
* lemma 1.1.4 , page 30 ) , there exists @xmath21 satisfying latexmath:[\[\label{3.3 } constant depending on @xmath2 .
let @xmath275 .
it is known that the sobolev space @xmath276 can be constructed by using fourier series .
precisely , we have @xmath277 @xmath276 is equipped with its natural norm
@xmath278 we take @xmath279 and we set @xmath280 in light of , we get @xmath281 here @xmath282 is a constant depending on @xmath2 .
therefore @xmath283 hence @xmath284 by cauchy - schwarz s inequality .
but @xmath285 this estimate in gives @xmath286 here @xmath287 .
then a consequence of theorem [ theorem2.2 ] is [ theorem3.1 ] let @xmath151 a measurable set of @xmath152 $ ] of positive lebesgue measure and @xmath138 .
there exists a constant @xmath129 , that can depend only on @xmath151 , @xmath2 and @xmath174 , so that , for any @xmath288 , @xmath289 we extend the previous result to a fractional one dimensional heat equation . to this end , for @xmath125 , we define @xmath290 , the fractional power of the operator @xmath291 under dirichlet boundary condition , as follows @xmath292 the ibvp for the heat equation for the fractional one dimensional laplacian is represented by the cauchy problem @xmath293 the solution of this cauchy problem is given by , where @xmath294 , @xmath295 if @xmath296 , we can apply again theorem [ theorem2.2 ] .
we get [ theorem3.1bis ] we assume @xmath296
. let @xmath151 a lebesgue - measurable set of @xmath152 $ ] of positive lebesgue measure and @xmath138 .
there exists a constant @xmath129 , that can depend only on @xmath151 , @xmath2 , @xmath297 and @xmath174 , so that , for any @xmath288 , @xmath298 next we consider the case @xmath299 . since @xmath300 , theorem [ theorem3.1bis ] is no longer valid in the present case .
we are going to apply theorem [ theorem2.4 ] instead of theorem [ theorem2.2 ] .
we pick @xmath301 for some @xmath179 .
let @xmath302 in light of , we get by using cauchy - schwarz s inequality @xmath303 but @xmath304 this and the first inequality in imply @xmath305 here and henceforth @xmath306 is a constant that can depend only on @xmath307 .
now a combination of and entails @xmath308 as @xmath309 is continuously embedded in @xmath310)$ ] , @xmath311 hence , it follows from and that @xmath312 similarly to , we prove @xmath313 thus @xmath314 by .
on the other hand , we have from @xmath315 in light of , and , we obtain by applying theorem [ theorem2.4 ] [ theorem3.2 ] let @xmath138 and @xmath179 .
there exists a constants @xmath129 , that can depend only on @xmath175 , @xmath297 , @xmath2 and @xmath174 , so that , for any @xmath316 , @xmath317 here @xmath318 is the constant in .
we observe that is satisfied for the sequence @xmath319 when @xmath320 .
the case @xmath321 is obvious and for the case @xmath322 it is a consequence of the following elementary inequality @xmath323 we mention that is possible to get a hlder stability estimate even when @xmath320 . to do that , we apply theorem [ theorem2.5 ] instead of theorem [ theorem2.4 ] .
we leave to the interested reader to write down the details .
let @xmath125 .
we recall that the solution of the fractional heat equation is given by , where @xmath294 , @xmath324 since @xmath325 we get by applying parseval s inequality @xmath326 on the other hand , @xmath327 when @xmath328 , @xmath329 , the biorthognal set in @xmath330 to @xmath331 , satisfies , for some constant @xmath129 depending on @xmath297 , @xmath332 this inequality is obtained from ( * ? ? ?
* estimate @xmath127 ) .
therefore we have , similarly to , @xmath333 and then @xmath334 estimate implies the following observability inequality @xmath335 let @xmath151 a lebesgue - measurable set of @xmath336 of positive lebesgue measure . as @xmath337 is given by a dirichlet series , we get from corollary [ corollary2.1 ] @xmath338 under the condition that @xmath339 .
let @xmath340 be the @xmath18-th fourier coefficient of @xmath341 .
then @xmath342 hence , for any @xmath343 , @xmath344 in light of , this estimate yields @xmath345 assuming in addition that @xmath346 , for some @xmath347 , we get @xmath348 here we used the fact that if @xmath349 , with @xmath347 , then @xmath339 . as before ,
this estimate allows us to prove the following theorem .
[ theorem4.1 ] let @xmath151 a lebesgue - measurable set of @xmath152 $ ] of positive lebesgue measure , @xmath347 and @xmath138 .
there exists a constant @xmath129 , that can depend only on @xmath151 , @xmath297 , @xmath269 and @xmath174 , so that , for any @xmath350 , @xmath351 we observe that if instead of we use , then we get a variant of theorem [ theorem4.1 ] in which is substituted by @xmath352 without the restriction that @xmath347 .
we have only to assume that @xmath37 .
let @xmath353 , where the sequence @xmath354 is non - resonant . from (
* proposition 5 ) the dirichlet - laplacian on @xmath38 has simple eigenvalues @xmath355 to each @xmath356 corresponds the eigenfunction @xmath357 so that @xmath358 forms an orthonormal basis of @xmath359 . to reduce the multidimensional case to the one dimensional case
, we need to restrict the initial sources to those of the form @xmath366 . in that case , @xmath367 where @xmath368 in other words , @xmath369 here @xmath370 is the solution of the one dimensional fractional heat equation when @xmath371 is substituted by @xmath372 . according to the parabolic maximum principle
( see for instance @xcite ) , we have @xmath373 let us assume that @xmath374 henceforth , @xmath2 and @xmath282 are the same as in . in light of , and , we get @xmath375 and then @xmath376 we fix @xmath138 and @xmath377 .
we prove similarly to theorem [ theorem3.1 ] that there exists a constant @xmath129 , that can depend only on @xmath174 and @xmath297 , so that , for any @xmath378 , @xmath379 , @xmath380 if @xmath381 is sufficiently small . hence , in light of , there is @xmath382 such that @xmath383 as we said in the introduction , there are only few results in the literature dealing with the problem of determining the initial heat distribution in a multidimensional heat equation from an overdermined data .
the usual overspecified data consists in an internal or a boundary measurement . we describe and comment briefly the main existing results and show the possible improvements .
let @xmath38 be a bounded domain of @xmath7 with @xmath388-smooth boundary @xmath389 .
let @xmath390 be the sequence of eigenvalues , counted according to their multiplicity , of the unbounded operator defined on @xmath359 by @xmath361 and @xmath391 .
let @xmath392 the corresponding sequence of eigenfunctions , chosen so that it forms an orthonormal basis of @xmath359 .
[ theorem1.1 ] let @xmath408 and @xmath138 .
then there exists a constant @xmath129 , depending on @xmath38 , @xmath397 ( resp .
@xmath398 ) , @xmath175 and @xmath174 , so that , for any @xmath409 , @xmath410 and @xmath411 [ remark4.1 ] 1 ) according to ( * ? ? ?
* theorem 1 ) , we can replace in , @xmath412 by @xmath413 , where @xmath76 is any lebsegue - measurable set contained in @xmath414 , having a non zero lebesgue measure .
we can also improve the estimate when @xmath415 contains a real - analytic open sub - manifold , that we denote by @xmath416 . in light of (
* theorem 2 ) , estimate holds true if @xmath417 is substituted by any lebesgue - measurable subset of @xmath418 with non zero lebesgue measure .
clearly , the mapping @xmath427 defines a norm on @xmath359 , weaker than the usual norm on @xmath359 .
therefore , can be reinterpreted as a lipschitz stability estimate of determining @xmath19 from @xmath428 : @xmath429 \3 ) let us show that , in the case of an internal measurement , we can directly get a stability estimate without using the observability inequality .
the key of this direct proof relies on lebeau - robbiano type inequality for the eigenfunctions @xmath431 . for @xmath432
, we set @xmath433 let @xmath398 be a lebesgue - measurable subset of @xmath38 of positive lebesgue measure . from ( * ? ? ?
* theorem 5 ) , we have @xmath434 but @xmath435 hence @xmath436.\ ] ] we integrate , with respect to @xmath437 , between @xmath438 and @xmath153 .
we get in a straightforward manner that @xmath439,\ ] ] implying that @xmath440 \\ & \le ce^{cn^{2/d}}\left[\int_{q_\omega}|u_f|^2dx+\sum_{n\geq n+1}|\widehat{f}_n|^2\right].\end{aligned}\ ] ] therefore , under the assumption @xmath441 for some @xmath186 , @xmath138 and @xmath442 , we obtain similarly to theorem [ theorem2.5 ] the following hlder stability estimate @xmath443 we end this subsection by mentioning that a lipschitz stability estimate was established in @xcite when the space of the initial heat distribution is given by a banach space , that we denote by @xmath444 in the sequel , built on the bergman - selberg space @xmath445 , where @xmath446 is a parameter . | we consider the problem of determining the initial heat distribution in the heat equation from a point measurement . we show that this inverse problem is naturally related to the one of recovering the coefficients of dirichlet series from its sum . taking the advantage of existing literature on dirichlet series , in connection with mntz s theorem ,
we establish various stability estimates of hlder and logarithmic type .
these stability estimates are then used to derive the corresponding ones for the original inverse problem , mainly in the case of one space dimension .
+ in higher space dimensions , we are interested to an internal or a boundary measurement . this issue is closely related to the problem of observability arising in control theory .
we complete and improve the existing results .
* keywords * : heat equation , fractional heat equation , initial heat distribution , mntz s theorem , point measurement , boundary measurement , stability estimates of hlder and logarithmic type .
* msc 2010 * : 35r30 |
models of strongly correlated electrons in two dimensions have been much studied in recent years as part of a search for a minimal model of the high - t@xmath6 superconducting cuprates , and other fascinating new antiferromagnetic systems@xcite .
the @xmath0 model , defined by the hamiltonian @xmath7 has become a generic model for mobile electrons / holes in doped antiferromagnets . in this expression the @xmath8 , @xmath9 are the usual electron creation and destruction operators , @xmath10 are the electron spin operators and the sum is over nearest neighbors pairs .
@xmath11 is a projection operator which ensures that doubly occupied states are excluded .
the @xmath0 model originally was derived@xcite as the strong correlation limit of the hubbard model , with @xmath12 , but it is now generally treated as an effective hamiltonian in its own right , with @xmath13 and @xmath14 independent parameters . for the cuprates
it is estimated@xcite that @xmath15 , which is well outside the region of validity of the hubbard mapping .
an independent `` derivation '' of the @xmath0 model as an effective hamiltonian for the cuprates has been presented by zhang and rice@xcite . at exactly half - filling ( one electron per site ) the kinetic term vanishes because of the single - occupancy restriction and the model reduces to an antiferromagnetic insulator .
it is known from a variety of numerical and analytic calculations@xcite that this model , on the square lattice , has nonzero long range antiferromagnetic ( nel ) order , reduced by strong quantum fluctuations .
removal of a small number of electrons ( by doping ) will allow mobility of holes and will reduce the antiferromagnetic order .
the physics of a small number of holes in a dynamic antiferromagnetic background remains a challenging problem .
previous studies of the @xmath0 model at @xmath1 have used a variety of approaches .
exact diagonalization methods have been used extensively on clusters up to 32 sites@xcite , but suffer from substantial finite - size corrections and an inability to treat extended excitations .
green s function monte carlo calculations have been performed for larger lattices , up to @xmath16@xcite , and show , for example , a significant decrease of the two - hole binding energy with increasing lattice size .
a variety of variational approaches and analytic many body methods have also been employed@xcite .
the picture that emerges from this body of work is roughly as follows . for one hole
the ground state energy is given by @xmath17 where the coefficients are approximate numerical estimates .
the dispersion curve for 1-hole states is qualitatively similar from different methods , with minima at @xmath18 and equivalent points .
the bandwidth from the largest diagonalizations scales as @xmath19 where the coefficients are again estimates and are not known with great precision .
the spectral function for 1-hole states has been calculated@xcite and shows a clear quasiparticle peak , together with a continuum at higher energies . for two holes
the binding energy , spectral function and pair susceptibility have been computed .
the results support the existence of a bound state with @xmath20 and d - wave symmetry for @xmath21 , with @xmath22 although not all methods agree and this is still controversial .
the dispersion curve for two - hole bound states is also of interest @xmath23 to our knowledge the only calculation of this is due to eder@xcite .
the aim of the present work is to study the physics of one - hole and two - hole states in the half - filled @xmath0 model at zero temperature via linked - cluster series expansions .
this approach has been developed extensively in recent years by our group@xcite and others@xcite .
most of the applications to date have been to spin systems , although we have included fermions in lattice gauge calculations .
shi and singh@xcite have used this approach to study the hubbard model at half filling , via an expansion about an additional ising term included in the hamiltonian . for completeness
we also mention the use of high - temperature expansions to study the @xmath0 model@xcite .
the present work of course addresses different questions which can not be probed by high - temperature series .
the plan of the paper is as follows . in sec .
ii , the series expansion method is briefly reviewed , and the results for one - hole properties of the @xmath0 model are presented . in sec .
iii , we study the dispersion relation and the binding energy for two - hole bound states . in sec .
iv , we study the one - hole and two - hole properties of the @xmath5 model .
v is devoted to a summary and discussion .
the series expansion method is based on a linked cluster formalism associated with standard rayleigh - schrdinger perturbation theory .
the ground state energy @xmath24 for a lattice of @xmath25 sites can be written as a sum over all connected clusters @xmath26 as @xmath27 where @xmath28 is a geometrical embedding factor and @xmath29 is the `` cumulant energy '' of cluster @xmath30 .
the same formula can be used for the energy of any particular cluster @xmath31 @xmath32 where the sum is now over all sub - clusters embeddable in @xmath31 .
the cluster energy is calculated perturbatively , then the cumulant energies are obtained iteratively from eq .
( [ eq3 ] ) and substituted into eq .
( [ eq2 ] ) to obtain the final series .
the formalism for excited states , needed to compute the dispersion curves , is a little more complex@xcite .
we introduce an ising spin anisotropy @xmath33\end{aligned}\ ] ] where @xmath34 , and rewrite the hamiltonian for the @xmath0 model as @xmath35 with @xmath36 with a staggered @xmath37-field of strength @xmath38 to improve convergence@xcite .
this term vanishes in the total hamiltonian in the isotropic exchange limit @xmath39 .
the ising part of the exchange , @xmath40 , forms the unperturbed hamiltonian while the perturbation @xmath41 includes both the hole hopping terms and transverse spin fluctuations .
the original hamiltonian for the @xmath0 model is recovered in the limit @xmath42 ( i.e. @xmath43 ) .
we expect the series expansion to give more reliable results in the small @xmath2 region , where the perturbation @xmath41 is less important .
we directly compute the series for the ground state energy difference between the system with one ( or two ) hole(s ) and the system with no hole .
two types of series can be obtained in this way . by fixing @xmath44 and @xmath38
we can compute series in the single variable @xmath45 up to order @xmath46 ( where @xmath46 is typically of order 11 - 13 ) , @xmath47 we refer to these as @xmath45-series .
extrapolation to @xmath42 is carried out by standard pad approximant or integrated differential approximant methods@xcite . in the other approach
we keep both @xmath45 and @xmath48 as expansion parameters and obtain double series of the form @xmath49 where the coefficients @xmath50 are computed up to order @xmath51 .
extrapolation to @xmath42 then yields a series in @xmath2 , @xmath52 which we refer to as a @xmath13-series . we first consider states with one hole , which may exist either on the up - sublattice ( a ) or the down - sublattice ( b ) , where the up and down sublattice are represented in fig . [ fig_lattice](a ) by open and full circles , respectively . here
@xmath53 are the usual primitive translation vectors .
the unperturbed wave function for the zero - momentum spin - up state is taken to be @xmath54 where @xmath55 is the classical nel state .
now a hole in the up ( down ) sublattice can never transform into a single hole in the down ( up ) sublattice , since the hamiltonian conserves the total spin @xmath56 ; so that we should choose @xmath57 shown in fig .
[ fig_lattice](a ) as the primitive translation vectors , rather than @xmath58 .
the unit vectors @xmath57 are those of the antiferromagnetic bipartite lattice .
the brillouin zones in reciprocal space corresponding to these different choices are shown in fig . [
fig_lattice](b ) .
most previous calculations , however , have used @xmath53 as unit vectors ( corresponding to @xmath59 in momentum space ) , instead of @xmath60 , so we will also conform to this convention ( and set @xmath61 to one ) unless it is specified otherwise . for momentum
@xmath62 we obtain the energy to order @xmath63 , involving 8739 distinct clusters containing up to 13 sites . for @xmath64 it is necessary to distinguish between clusters having different spatial orientation .
we expand the energy to order @xmath65 , involving 23546 clusters with up to 11 sites .
the series coefficients for @xmath66 at fixed momentum @xmath67 and different @xmath38 are listed in table [ tab_1hmin ] , where we can see the series converges best around @xmath68 .
the extrapolations of these series by using the integrated differential approximant methods are shown in figure [ fig_diffr ] , where we can see that at @xmath42 , the extrapolations from different @xmath38 series agree with each other very well , while the case @xmath69 gives the smallest error bar . in table [ tab_tjk ] , we list the full spectrum series for @xmath66 and @xmath70 . the other coefficients are too lengthy to publish but can be supplied on request . in figure [ fig_ek1h ]
we show the energy of a one - hole state for @xmath71 along symmetry directions in the brillouin zone , for various ratios @xmath2 obtained from @xmath45-series .
the lowest energy occurs at @xmath72 and equivalent points .
this result , and the overall shape of the one - hole dispersion curve are in very good agreement with previous work@xcite .
we note that the symmetry and the redundant information evident in fig .
[ fig_ek1h ] are the results of choosing a brillouin zone which is two times too big .
if we choose @xmath73 as defining the bipartite brillouin zone , the corresponding dispersion relation along symmetry directions in the brillouin zone is shown in fig .
[ fig_ek1h2 ] .
figure [ fig_w ] shows the one - hole bandwidth , @xmath74 , as a function of @xmath75 , compared with results from other approaches . for @xmath75 around 0.5 our results
agree well with other work , but for large @xmath75 our result is significantly larger than that obtained by martinez and horsch@xcite .
also we found a peak value of @xmath76 at @xmath77 , rather than @xmath78 at @xmath79 as obtained by martinez and horsch@xcite , and liu and manousakis@xcite . using the two - variable expansion
we obtain the one - hole energy @xmath80 at some particular values of @xmath71 : @xmath81 \nonumber \\
e^{\rm 1h}(\pi/2,\pi/2)/j & = & 2.164(2 ) -3.91(4 ) ( t / j)^2 + 9.1(4 ) ( t / j)^4 + o[(t / j)^6 ] \label{eq_e1h } \\
e^{\rm 1h}(\pi,0)/j & = & 2.164(2 ) -2.68(2 ) ( t / j)^2 + 3.06(4 ) ( t / j)^4 + o[(t / j)^6]\nonumber % old res .
from y=0 only % e^{\rm 1h } ( 0,0)/j & = & 2.164(3 ) -0.650(6 ) ( t / j)^2 - 0.036(9 ) ( t / j)^4 + o[(t / j)^6 ] \nonumber \\ % e^{\rm 1h}(\pi/2,\pi/2)/j & = & 2.164(3 ) -3.91(4 ) ( t / j)^2 + 9.1(4 ) ( t / j)^4 + o[(t / j)^6 ] \label{eq_e1h } \\
% e^{\rm 1h}(\pi,0)/j & = & 2.164(3 ) -2.70(4 ) ( t / j)^2 + 3.01(4 ) ( t / j)^4 + o[(t / j)^6]\nonumber\end{aligned}\ ] ] these coefficients are not exact as they involve an extrapolation to @xmath42 . the leading coefficient 2.164(3 )
can be compared with 2.1552 from a spin wave calculation@xcite and 2.193(7 ) from a monte carlo simulation@xcite .
the bandwidth can be obtained as @xmath82 \label{w_2p}\ ] ] this result is also plotted in figure [ fig_w ] as a dashed line , and it agrees well at small @xmath2 with that obtained from @xmath45-series . the leading coefficient 3.26(5 )
should be compared with previous calculations giving a value of 2@xcite or 2.6@xcite .
in the previous section we developed a series expansion approach to calculate one - hole properties of the @xmath0 model , and showed how reliable quantitative results could be obtained . in the present section we extend this approach to consider states with two holes .
we start from an unperturbed wave - function consisting of two nearest - neighbour holes with opposite spin , and overall momentum @xmath71 : @xmath83 here again there are two different choices of brillouin zone possible .
one may choose to use a diagonal unit vector of length @xmath84 ( where @xmath61 is the lattice spacing ) , which connects the centre points of two adjacent bonds of the original lattice as shown in fig .
[ fig_lattice2h](a ) ; this corresponds to the large , outer brillouin zone in fig . [ fig_lattice2h](b ) .
this was basically the convention used by eder@xcite . alternatively , and more correctly
, one may choose a diagonal unit vector of length @xmath85 , differentiating between the @xmath86 and @xmath87 sublattices .
this gives rise again to the brillouin zone of the bipartite lattice , marked as the inner zone in fig .
[ fig_lattice2h](b ) . in this second case
, there are 4 independent bond states in the unit cell , and one may form linear combinations of them making up the conventional @xmath88- , @xmath89- , @xmath90- and @xmath3-wave states . with the first choice , all these four states are included within one brillouin zone : the points corresponding to zero - momentum @xmath88- , @xmath89- , @xmath90- and @xmath3-wave states are shown- '' and `` @xmath90- '' wave states are degenerate by reflection symmetry .
the points marked @xmath89 and @xmath90 in fig .
[ fig_lattice2h](b ) are really linear combinations of the proper @xmath89 and @xmath90 states .
[ fig_lattice2h](b ) .
then one finds that the different states are related by equations such as @xmath91 ( in units where @xmath84 has been rescaled to one ) .
this relationship is evident in fig .
19 of ref .
@xcite .
we have computed series for the two - hole spectrum @xmath92 up to order @xmath65 .
this involves 46,440 distinct clusters of up to 12 sites .
the @xmath45-series coefficients for @xmath66 and @xmath70 are listed in table [ tab_tjk ] . using the two - variable expansion , after extrapolation to @xmath42 , we obtain the following estimates for the two - hole @xmath88- , @xmath4- and @xmath3-wave pair energy : @xmath93\nonumber \\ e^{2p}/j & \equiv & e^{\rm 2h}(0,\pi)/j = 3.798(2 ) - 3.705(10 ) ( t / j)^2 - 4.1(1 ) ( t / j)^4 + o[(t / j)^6 ] \label{eq_e2h } \\ e^{2s}/j & \equiv & e^{\rm 2h}(\pi,\pi)/j = 3.798(2 ) - 0.52(3 ) ( t / j)^2 - 1.62(5 ) ( t / j)^4 + o[(t / j)^6 ] \nonumber \end{aligned}\ ] ]
so the binding energies defined by @xmath94 for @xmath88- , @xmath4- and @xmath3-wave pair states are : @xmath95 \nonumber \\ e_{\rm b}^{2p}/j & = & -0.529(4 ) + 4.1(1 ) ( t / j)^2 - 22.3(9 ) ( t / j)^4 + o[(t / j)^6 ] \label{eq_binding } \\ e_{\rm b}^{2s}/j & = & -0.529(4 ) + 7.3(1 ) (
t / j)^2 - 19.8(9 ) ( t / j)^4 + o[(t / j)^6 ] \nonumber\end{aligned}\ ] ] the binding energy for @xmath96 has previously been given as @xmath97 by spin wave theory@xcite , and @xmath98 by monte carlo@xcite .
our result is in relatively good agreement but suggests that the monte carlo calculations may overestimate the binding energy somewhat .
we have also computed the dispersion curves for the two - hole bound state from @xmath45-series , and these are shown in figure [ fig_ek2h ] , for various small @xmath2 .
the minimum occurs at @xmath99 ( @xmath3-wave pair ) , and the maximum occurs at @xmath100 ( @xmath88-wave pair ) , while the energy at @xmath100 is almost unchanged as @xmath2 increases .
the curves are quite flat for small @xmath2 , as expected , and develop more structure as @xmath2 increases .
the results are broadly similar to those of eder@xcite .
the @xmath88- , @xmath4- and @xmath3-wave pair binding energies , obtained from both the @xmath45-series and the two - variable expansion as a function of @xmath2 , are shown in figure [ fig_binding ] .
note that for @xmath45-series , the extrapolation is done on series for @xmath101 , rather than the 1-hole and 2-hole energies individually , because values of these two energies are quite large , while their difference is usually small .
the convergence is excellent for small @xmath2 , and we believe our results are very accurate there .
the uncertainty increases for larger @xmath13 .
we can see from this figure that as @xmath2 increases from zero , the @xmath88-wave binding energy decreases very quickly , and the binding disappears around @xmath102 .
the @xmath4-wave binding energy also decreases quickly in the small @xmath2 region , but it turns out to be very flat beyond @xmath103 , and the binding energy is very small .
the @xmath3-wave binding energy decreases slightly as @xmath2 increases from zero .
our results are consistent with the @xmath3-wave binding disappearing around @xmath104 .
beyond that point , the @xmath4-wave state appears to have the lower energy , and remains very weakly bound according to our results .
figure [ fig_binding ] also shows some other numerical results for the 2-hole binding energy .
the solid circles with error bars are gfmc results for the @xmath16 lattice obtained by boninsegni and manousakis@xcite , and lie somewhat below our result , though broadly similar in trend : the discrepancy may easily be attributed to a finite - size effect . the solid triangles are the results of a recent 32-site finite - lattice calculation@xcite .
the open circles are the results of a linear extrapolation in @xmath105 from the finite - lattice data by chernyshev , leung and gooding@xcite ; but they discuss reasons why such a naive extrapolation is not to be relied on especially at large @xmath2 .
the results lie somewhat above our series results for the @xmath3-wave binding energy , but again have a broadly similar trend .
chernyshev , leung and gooding@xcite find that the lowest bound state on the 32-site lattice is a singlet state with @xmath106 symmetry .
the apparent cross - over between the @xmath4-wave and @xmath3-wave bound states came initially as a surprise .
but in fact it accords very well with predictions@xcite based on the kohn - luttinger effect . in a repulsive fermi liquid ,
pairing in high - angular momentum states can occur due to a singularity in the 2-particle scattering amplitude .
this effect was previously believed not to occur in two dimensions , because it does not appear up to second order in perturbation theory ; but it was recently shown@xcite that a singularity does arise at higher order .
chubukov@xcite showed that in the dilute limit the effect survives down to @xmath107 and gives rise to @xmath4-wave pairing .
he also pointed out that the same effect will occur in an arbitrarily dense fermi liquid .
both numerical@xcite and analytical@xcite approaches have shown that at low densities @xmath4-wave pairing is dominant for @xmath108 , @xmath3-wave is dominant at @xmath109 , and there is a region of @xmath88-wave dominance for @xmath110 . figure [ fig_binding ] appears to indicate that a similar pattern extends even to the region of half - filling .
in fact , the earlier variational calculation in the hamiltonian .
] of eder also shows a crossover between @xmath4-wave and @xmath3-wave states ( fig .
21 of ref .
@xcite ) at around @xmath111 .
several theoretical approaches@xcite have predicted that magnon exchange should give rise to shallow , long - range bound states even at large @xmath2 values .
the behaviour of the @xmath4-wave state looks consistent with this idea , but not the @xmath3-wave state .
it is possible that another state ( or states ) in the @xmath3-wave sector could cross over the one we have calculated , and become the lowest energy state at large @xmath2 .
such a crossover could typically imply a square - root singularity in the energy eigenvalue , and the series extrapolation would no longer be valid beyond that point .
we also study a simplified version of the @xmath0 model , the @xmath5 model , which has an ising rather than a heisenberg spin interaction @xmath112 here we naturally take the first term in @xmath113 as the unperturbed hamiltonian , and the second term as the perturbation , and perform an expansion in @xmath48 .
this expansion only converges well up to about @xmath114 . to improve the convergence in the large @xmath48 region
, we can include a staggered @xmath37-field of strength @xmath38 as in the @xmath115 model , and perform the following separation , @xmath116 where @xmath117 we then take @xmath40 as the unperturbed hamiltonian and @xmath41 as the perturbation , and perform an expansion in @xmath45 for fixed values of @xmath48 and @xmath38 . here
again we need to extrapolate the series to @xmath42 .
the series for the one - hole state have been computed up to order 21 ( that is , to @xmath118 for the hamiltonian in eq .
( [ h_tjz ] ) or to order @xmath119 for the hamiltonian in eq .
( [ h_tjz2 ] ) ) involving a list of 23546 clusters with up to 11 sites , while the series for the two - hole bound state have been computed up to order 18 , involving a list of 23546 clusters up to 11 sites .
the final series in @xmath48 for the spectrum of the one - hole state and two - hole bound state are listed in table [ tabtjz ] .
[ fig_ek1h_tjz ] shows the one - hole spectrum for some particular values of @xmath48 .
here we can see that the shape of the dispersion curve is quite different from that for the @xmath0 model : the lowest energy occurs at @xmath100 rather than @xmath18 as in the @xmath0 model , and the bandwidth here is about 10 times smaller than that for the @xmath0 model .
the minimum one - hole energy as a function of @xmath120 is shown in fig .
[ fig_e1hmin ] . in the interval @xmath121
, the minimum one - hole energy can be very well fitted@xcite as @xmath122 where the linear term comes from the @xmath123 term in the hamiltonian .
this is in good agreement with previous estimates@xcite .
the two - hole dispersion is shown in fig .
[ fig_ek2s_tjz ] .
we note that the energy at @xmath100 is exactly @xmath124 , independent of @xmath13 , and the energy along path @xmath125 ( @xmath4-wave pair ) to @xmath99 ( @xmath3-wave pair ) is very flat .
the @xmath88- , @xmath4- and @xmath3-wave binding energies defined by @xmath126 are shown in figure [ fig_binding_tjz ] .
we can see from this figure that @xmath88-wave binding disappears around @xmath127 .
the @xmath3-wave binding energies are slightly greater than for @xmath4-wave when @xmath128 , but when @xmath129 , the @xmath4-wave pair has a greater binding energy than the @xmath3-wave , similarly to that found for the full @xmath0 model .
thus the @xmath5 model also shows a crossover between @xmath3-wave and @xmath4-wave bound states , occurring even earlier than in the full @xmath0 model .
again , the earlier variational calculations of eder showed very similar results ( fig .
12 of ref .
@xcite ) .
the solid points in figure [ fig_binding_tjz ] are the results of a reduced basis diagonalization study on clusters of up to 50 sites by riera and dagotto@xcite .
the trend of these results agrees extremely well with the series estimates , although their estimates lie systematically a little lower .
they do not discuss the symmetry of the bound state@xcite .
they find the binding of the 2-hole state disappears at @xmath130 , or @xmath131 , which agrees very well with our results .
the @xmath0 model on the square lattice has been studied by many workers over the last decade , using a variety of techniques .
these methods are all approximate and have uncertain accuracy .
independent confirmation from a variety of approaches increases one s confidence in the results .
we have used the linked - cluster series expansion method to study one and two hole properties of the half - filled @xmath0 and @xmath5 models in 2-dimensions at @xmath1 . to our knowledge
this is the first time this has been done .
our results for 1-hole and 2-hole states can be summarized as follows : * the 1-hole dispersion curve has been computed throughout the brillouin zone , and agrees in shape with previous calculations . *
the minimum for 1-hole dispersion occurs at @xmath132 and equivalent points , again in agreement with previous work . *
the bandwidth @xmath133 for the 1-hole dispersion curve of the @xmath0 model has been calculated . at small @xmath75
it agrees with the spin - wave calculations of martinez and horsch@xcite and liu and manousakis@xcite , but beyond @xmath134 our results lie substantially higher than theirs .
we get a peak value of @xmath135 at @xmath136 , rather than @xmath78 at @xmath137 as given in previous work .
* we have computed the dispersion curve for 2-hole bound states . to our knowledge
this has not been well calculated previously . * in the limit @xmath138 of the @xmath0 model , we can compare our results for the one - hole and two - hole energy ( @xmath139 and @xmath140 ) , the binding energy ( @xmath141 ) , and the bandwidth for the one - hole state ( @xmath133 ) with previous spin - wave@xcite , monte carlo@xcite and other calculations .
+ [ cols="<,^,<",options="header " , ] rr|rr|rr|rr & & & & & & & + + ( 0 , 0 , 0 ) & 2.000000000 & ( 8 , 2 , 0 ) & -8.693161815@xmath142 & ( 5 , 1 , 1 ) & -3.650599647@xmath143 & ( 11 , 5 , 1 ) & -2.756493796@xmath142 + ( 2 , 0 , 0 ) & -5.333333333@xmath143 & ( 9 , 2 , 0 ) & 2.176985203@xmath143 & ( 6 , 1 , 1 ) & 2.714260859@xmath143 & ( 6 , 2 , 2 ) & 7.130158730@xmath142 + ( 4 , 0 , 0 ) & 1.529021164@xmath143 & ( 10 , 2 , 0 ) & -1.017391598@xmath142 & ( 7 , 1 , 1 ) & 2.890627842@xmath143 & ( 7 , 2 , 2 ) & -6.731669606@xmath142 + ( 5 , 0 , 0 ) & 9.481481481@xmath142 & ( 11 , 2 , 0 ) & -3.628184110@xmath143 & ( 8 , 1 , 1 ) & -2.000742064@xmath143 & ( 8 , 2 , 2 ) & -1.075223788@xmath143 + ( 6 , 0 , 0 ) & -1.772367448@xmath142 & ( 6 , 4 , 0 ) & 1.188359788@xmath142 & ( 9 , 1 , 1 ) & -1.603987396@xmath143 & ( 9 , 2 , 2 ) & 1.539341160@xmath143 + ( 7 , 0 , 0 ) & -1.684678469@xmath143 & ( 7 , 4 , 0 ) & -4.980599647@xmath144 & ( 10 , 1 , 1 ) & -1.153947213@xmath143 & ( 10 , 2 , 2 ) & 9.103499935@xmath142 + ( 8 , 0 , 0 ) & -6.427987242@xmath142 & ( 8 , 4 , 0 ) & -1.402790612@xmath142 & ( 11 , 1 , 1 ) & 1.409738327@xmath143 & ( 11 , 2 , 2 ) & -2.080643863@xmath143 + ( 9 , 0 , 0 ) & 2.538548842@xmath143 & ( 9 , 4 , 0 ) & 2.637981607@xmath142 & ( 6 , 3 , 1 ) & 9.506878307@xmath142 & ( 9 , 4 , 2 ) & 3.146104767@xmath142 + ( 10 , 0 , 0 ) & 1.029816377@xmath143 & ( 10 , 4 , 0 ) & -1.865761552@xmath142 & ( 7 , 3 , 1 ) & -7.727789536@xmath142 & ( 10 , 4 , 2 ) & -4.054217598@xmath142 + ( 11 , 0 , 0 ) & -2.786820323@xmath143 & ( 11 , 4 , 0 ) & -4.652610741@xmath142 & ( 8 , 3 , 1 ) & -1.457528700@xmath143 & ( 11 , 4 , 2 ) & -7.000652266@xmath142 + ( 3 , 2 , 0 ) & 2.666666667@xmath143 & ( 9 , 6 , 0 ) & 1.048701589@xmath144 & ( 9 , 3 , 1 ) & 1.806575878@xmath143 & ( 9 , 3 , 3 ) & 2.097403178@xmath142 + ( 4 , 2 , 0 ) & -5.333333333@xmath142 & ( 10 , 6 , 0 ) & -6.651097800@xmath145 & ( 10 , 3 , 1 ) & 8.928712117@xmath142 & ( 10 , 3 , 3 ) & -2.874385753@xmath142 + ( 5 , 2 , 0 ) & -1.173562610@xmath143 & ( 11 , 6 , 0 ) & -1.940149264@xmath144 & ( 11 , 3 , 1 ) & -2.245856051@xmath143 & ( 11 , 3 , 3 ) & -4.632188322@xmath142 + ( 6 , 2 , 0 ) & 8.604515263@xmath142 & ( 3 , 1 , 1 ) & 5.333333333@xmath143 & ( 9 , 5 , 1 ) & 1.258441907@xmath142 + ( 7 , 2 , 0 ) & -1.320134011@xmath142 & ( 4 , 1 , 1 ) & -2.962962963@xmath143 & ( 10 , 5 , 1 ) & -1.312853800@xmath142 + + ( 0 , 0 , 0 ) & 3.500000000 & ( 6 , 3 , 3 ) & 1.457965902@xmath144 & ( 6 , 3 , 2 ) & 5.831863610@xmath144 & ( 9 , 4 , 2 ) & @xmath237.128118497@xmath145 + ( 2 , 0 , 0 ) & @xmath237.333333333@xmath143 & ( 8 , 3 , 3 ) & @xmath238.399019298@xmath144 & ( 7 , 3 , 2 ) & @xmath234.705118313@xmath144 & ( 10 , 4 , 2 ) & @xmath236.158730192@xmath142 + ( 4 , 0 , 0 ) & 1.945310847@xmath143 & ( 9 , 3 , 3 ) & 2.106020257@xmath142 & ( 8 , 3 , 2 ) & @xmath235.937350737@xmath142 & ( 11 , 4 , 2 ) & 6.702194733@xmath142 + ( 5 , 0 , 0 ) & 1.570370370@xmath143 & ( 10 , 3 , 3 ) & 5.953632826@xmath144 & ( 9 , 3 , 2 ) & 3.295758313@xmath142 & ( 10 , 5 , 3 ) & 2.643901061@xmath145 + ( 6 , 0 , 0 ) & @xmath231.814429222@xmath143 & ( 11 , 3 , 3 ) & 4.246104998@xmath144 & ( 10 , 3 , 2 ) & 1.653086795@xmath142 & ( 11 , 5 , 3 ) & @xmath238.754870105@xmath146 + ( 7 , 0 , 0 ) & 3.336368382@xmath142 & ( 8 , 4 , 4 ) & 1.843920010@xmath145 & ( 11 , 3 , 2 ) & @xmath231.386696140@xmath142 & ( 6 , 3 , 0 ) & 5.831863610@xmath144 + ( 8 , 0 , 0 ) & @xmath231.977570791@xmath143 & ( 10 , 4 , 4 ) & @xmath233.073373856@xmath144 & ( 8 , 4 , 3 ) & 7.375680039@xmath145 & ( 7 , 3 , 0 ) & @xmath231.162496326@xmath144 + ( 9 , 0 , 0 ) & @xmath231.214962788@xmath143 & ( 11 , 4 , 4 ) & 4.127194476@xmath145 & ( 9 , 4 , 3 ) & @xmath235.394036567@xmath145 & ( 8 , 3 , 0 ) & @xmath231.040888381@xmath143 + ( 10 , 0 , 0 ) & 1.897994391@xmath143 & ( 10 , 5 , 5 ) & 2.643901061@xmath146 & ( 10 , 4 , 3 ) & @xmath232.344210680@xmath142 & ( 9 , 3 , 0 ) & 8.179624689@xmath142 + ( 11 , 0 , 0 ) & 7.710636344@xmath143 & ( 2 , 1 , 0 ) & 6.666666667@xmath143 & ( 11 , 4 , 3 ) & 3.300423401@xmath142 & ( 10 , 3 , 0 ) & 1.171660236@xmath143 + ( 2 , 1 , 1 ) & 3.333333333@xmath143 & ( 3 , 1 , 0 ) & 3.333333333@xmath143 & ( 10 , 5 , 4 ) & 1.057560424@xmath145 & ( 11 , 3 , 0 ) & 1.531781787@xmath142 + ( 3 , 1 , 1 ) & @xmath233.333333333@xmath143 & ( 4 , 1 , 0 ) & @xmath232.673809524@xmath143 & ( 11 , 5 , 4 ) & @xmath237.959003190@xmath146 & ( 8 , 4 , 1 ) & 2.212704012@xmath144 + ( 4 , 1 , 1 ) & @xmath231.192857143@xmath143 & ( 5 , 1 , 0 ) & @xmath234.024977954@xmath143 & ( 4 , 2 , 0 ) & 2.962962963@xmath142 & ( 9 , 4 , 1 ) & @xmath238.862200427@xmath145 + ( 5 , 1 , 1 ) & 2.904905203@xmath143 & ( 6 , 1 , 0 ) & @xmath231.302727805@xmath143 & ( 5 , 2 , 0 ) & @xmath232.629629630@xmath142 & ( 10 , 4 , 1 ) & @xmath239.978608234@xmath142 + ( 6 , 1 , 1 ) & 5.329329186@xmath143 & ( 7 , 1 , 0 ) & 3.499536691@xmath143 & ( 6 , 2 , 0 ) & @xmath232.375711537@xmath143 & ( 11 , 4 , 1 ) & 1.153622881@xmath143 + ( 7 , 1 , 1 ) & @xmath236.271803816@xmath143 & ( 8 , 1 , 0 ) & 4.079418136@xmath143 & ( 7 , 2 , 0 ) & 2.330090787@xmath143 & ( 10 , 5 , 2 ) & 4.230241698@xmath145 + ( 8 , 1 , 1 ) & @xmath237.522337850@xmath143 & ( 9 , 1 , 0 ) & 5.301735104@xmath142 & ( 8 , 2 , 0 ) & 3.039557383@xmath143 & ( 11 , 5 , 2 ) & @xmath231.750974021@xmath145 + ( 9 , 1 , 1 ) & 5.347837206@xmath143 & ( 10 , 1 , 0 ) & @xmath231.026434348 & ( 9 , 2 , 0 ) & @xmath233.529882617@xmath143 & ( 8 , 4 , 0 ) & 1.106352006@xmath144 + ( 10 , 1 , 1 ) & 1.146091910 & ( 11 , 1 , 0 ) & @xmath237.412817539@xmath143 & ( 10 , 2 , 0 ) & @xmath236.813379022@xmath143 & ( 9 , 4 , 0 ) & @xmath237.128118497@xmath145 + ( 11 , 1 , 1 ) & @xmath234.863394606@xmath143 & ( 4 , 2 , 1 ) & 5.925925926@xmath142 & ( 11 , 2 , 0 ) & 9.790814131@xmath143 & ( 10 , 4 , 0 ) & @xmath235.538838348@xmath142 + ( 4 , 2 , 2 ) & 1.481481481@xmath142 & ( 5 , 2 , 1 ) & @xmath232.629629630@xmath142 & ( 6 , 3 , 1 ) & 8.747795414@xmath144 & ( 11 , 4 , 0 ) & 8.043954590@xmath142 + ( 6 , 2 , 2 ) & @xmath235.533170820@xmath142 & ( 6 , 2 , 1 ) & @xmath238.610412317@xmath142 & ( 7 , 3 , 1 ) & @xmath235.867614638@xmath144 & ( 10 , 5 , 1 ) & 5.287802122@xmath145 + ( 7 , 2 , 2 ) & 1.147526740@xmath142 & ( 7 , 2 , 1 ) & 5.769980192@xmath142 & ( 8 , 3 , 1 ) & @xmath231.459267389@xmath143 & ( 11 , 5 , 1 ) & @xmath232.626461032@xmath145 + ( 8 , 2 , 2 ) & 3.484612337@xmath142 & ( 8 , 2 , 1 ) & 2.118965896@xmath143 & ( 9 , 3 , 1 ) & 1.611775400@xmath143 & ( 10 , 5 , 0 ) & 3.172681273@xmath145 + ( 9 , 2 , 2 ) & @xmath238.489781072@xmath142 & ( 9 , 2 , 1 ) & @xmath239.184491129@xmath142 & ( 10 , 3 , 1 ) & 1.623787831@xmath143 & ( 11 , 5 , 0 ) & @xmath239.550737020@xmath146 + ( 10 , 2 , 2 ) & @xmath238.365174384@xmath142 & ( 10 , 2 , 1 ) & @xmath233.638022235@xmath143 & ( 11 , 3 , 1 ) & @xmath233.687197934@xmath143 + ( 11 , 2 , 2 ) & 3.945274143@xmath143 & ( 11 , 2 , 1 ) & @xmath235.611998341@xmath144 & ( 8 , 4 , 2 ) & 1.475136008@xmath144 + rr|rr|rr|rr & & & & & & & + + ( 0 , 0 , 0 ) & 2.000000000 & ( 20 , 0 , 0 ) & 3.269830958@xmath147 & ( 8 , 1 , 1 ) & 1.107677837 & ( 20 , 3 , 1 ) & 1.363566623 + ( 2 , 0 , 0 ) & @xmath232.666666667 & ( 10 , 2 , 0 ) & @xmath239.187011686@xmath144 & ( 10 , 1 , 1 ) & @xmath235.258263841 & ( 12 , 2 , 2 ) & 2.213826787@xmath142 + ( 4 , 0 , 0 ) & 2.607407407 & ( 12 , 2 , 0 ) & 1.237459977@xmath143 & ( 12 , 1 , 1 ) & 2.361854957@xmath148 & ( 14 , 2 , 2 ) & @xmath232.801396443@xmath143 + ( 6 , 0 , 0 ) & @xmath235.667818930 & ( 14 , 2 , 0 ) & @xmath231.042288265 & ( 14 , 1 , 1 ) & @xmath231.041949744@xmath149 & ( 16 , 2 , 2 ) & 2.293697931 + ( 8 , 0 , 0 ) & 1.579276071@xmath148 & ( 16 , 2 , 0 ) & 7.139692772 & ( 16 , 1 , 1 ) & 4.570753372@xmath149 & ( 18 , 2 , 2 ) & @xmath231.550982309@xmath148 + ( 10 , 0 , 0 ) & @xmath234.977461910@xmath148 & ( 18 , 2 , 0 ) & @xmath234.367374242@xmath148 & ( 18 , 1 , 1 ) & @xmath232.003192131@xmath150 & ( 20 , 2 , 2 ) & 9.435217647@xmath148 + ( 12 , 0 , 0 ) & 1.689229038@xmath149 & ( 20 , 2 , 0 ) & 2.492266705@xmath149 & ( 20 , 1 , 1 ) & 8.787977137@xmath150 & ( 20 , 4 , 2 ) & 4.053738087@xmath146 + ( 14 , 0 , 0 ) & @xmath236.023423934@xmath149 & ( 18 , 4 , 0 ) & @xmath231.275445883@xmath151 & ( 14 , 3 , 1 ) & @xmath231.112602736@xmath145 & ( 18 , 3 , 3 ) & @xmath234.160122546@xmath144 + ( 16 , 0 , 0 ) & 2.225264150@xmath150 & ( 20 , 4 , 0 ) & 1.432818794@xmath145 & ( 16 , 3 , 1 ) & 6.891013162@xmath144 & ( 20 , 3 , 3 ) & 8.089745849@xmath142 + ( 18 , 0 , 0 ) & @xmath238.447437043@xmath150 & ( 6 , 1 , 1 ) & @xmath231.896296296@xmath143 & ( 18 , 3 , 1 ) & @xmath231.208165780@xmath143 + + ( 0 , 0 , 0 ) & 3.500000000 & ( 2 , 1 , 0 ) & 2.666666667 & ( 10 , 3 , 1 ) & @xmath233.833568660 & ( 10 , 4 , 0 ) & @xmath234.187749256@xmath143 + ( 2 , 0 , 0 ) & @xmath234.000000000 & ( 4 , 1 , 0 ) & @xmath232.607407407 & ( 12 , 3 , 1 ) & 3.749428458@xmath148 & ( 12 , 4 , 0 ) & @xmath234.625682644 + ( 4 , 0 , 0 ) & 2.311111111 & ( 6 , 1 , 0 ) & @xmath239.124044680@xmath143 & ( 14 , 3 , 1 ) & @xmath231.763568623@xmath149 & ( 14 , 4 , 0 ) & 3.879009243@xmath148 + ( 6 , 0 , 0 ) & @xmath233.578130511@xmath143 & ( 8 , 1 , 0 ) & 2.709462265@xmath148 & ( 16 , 3 , 1 ) & 5.362296757@xmath149 & ( 16 , 4 , 0 ) & @xmath231.755530196@xmath149 + ( 8 , 0 , 0 ) & @xmath231.447791909@xmath148 & ( 10 , 1 , 0 ) & @xmath231.454354060@xmath149 & ( 18 , 3 , 1 ) & @xmath234.338040480@xmath149 & ( 18 , 4 , 0 ) & 5.141129944@xmath149 + ( 10 , 0 , 0 ) & 8.449051521@xmath148 & ( 12 , 1 , 0 ) & 4.973154521@xmath149 & ( 8 , 4 , 2 ) & 3.776348180@xmath143 & ( 10 , 5 , 1 ) & 5.414709373@xmath143 + ( 12 , 0 , 0 ) & @xmath233.205718341@xmath149 & ( 14 , 1 , 0 ) & @xmath238.044807264@xmath149 & ( 10 , 4 , 2 ) & @xmath237.694903208@xmath143 & ( 12 , 5 , 1 ) & @xmath236.844512414@xmath143 + ( 14 , 0 , 0 ) & 7.683946158@xmath149 & ( 16 , 1 , 0 ) & @xmath233.592648436@xmath150 & ( 12 , 4 , 2 ) & @xmath233.834200865 & ( 14 , 5 , 1 ) & @xmath231.368955321@xmath148 + ( 16 , 0 , 0 ) & 1.864995221@xmath149 & ( 18 , 1 , 0 ) & 4.011882196@xmath147 & ( 14 , 4 , 2 ) & 4.301073460@xmath148 & ( 16 , 5 , 1 ) & 1.169642314@xmath149 + ( 18 , 0 , 0 ) & @xmath231.447204343@xmath147 & ( 4 , 2 , 1 ) & 9.481481481@xmath143 & ( 16 , 4 , 2 ) & @xmath232.363456039@xmath149 & ( 18 , 5 , 1 ) & @xmath235.467419040@xmath149 + ( 2 , 1 , 1 ) & 1.333333333 & ( 6 , 2 , 1 ) & @xmath232.072004703 & ( 18 , 4 , 2 ) & 9.094821992@xmath149 & ( 12 , 6 , 2 ) & 5.024245295@xmath143 + ( 4 , 1 , 1 ) & @xmath231.362962963 & ( 8 , 2 , 1 ) & 1.071796284 & ( 10 , 5 , 3 ) & 2.707354687@xmath143 & ( 14 , 6 , 2 ) & @xmath235.607740036@xmath143 + ( 6 , 1 , 1 ) & 3.117601411 & ( 10 , 2 , 1 ) & 1.934131152@xmath148 & ( 12 , 5 , 3 ) & @xmath236.099856791@xmath143 & ( 16 , 6 , 2 ) & @xmath231.688448287@xmath148 + ( 8 , 1 , 1 ) & @xmath231.157205899@xmath148 & ( 12 , 2 , 1 ) & @xmath231.610227509@xmath149 & ( 14 , 5 , 3 ) & @xmath233.591949866 & ( 18 , 6 , 2 ) & 1.530376091@xmath149 + ( 10 , 1 , 1 ) & 3.724437583@xmath148 & ( 14 , 2 , 1 ) & 9.100666397@xmath149 & ( 16 , 5 , 3 ) & 4.680082984@xmath148 & ( 14 , 7 , 3 ) & 4.576926174@xmath143 + ( 12 , 1 , 1 ) & @xmath234.436033620@xmath148 & ( 16 , 2 , 1 ) & @xmath234.053430224@xmath150 & ( 18 , 5 , 3 ) & @xmath232.944841481@xmath149 & ( 16 , 7 , 3 ) & @xmath234.694339733@xmath143 + ( 14 , 1 , 1 ) & @xmath235.334708723@xmath149 & ( 18 , 2 , 1 ) & 1.279544679@xmath147 & ( 12 , 6 , 4 ) & 2.009698118@xmath143 & ( 18 , 7 , 3 ) & @xmath231.886986003@xmath148 + ( 16 , 1 , 1 ) & 5.421019234@xmath150 & ( 6 , 3 , 2 ) & 3.732392710@xmath143 & ( 14 , 6 , 4 ) & @xmath234.925030503@xmath143 & ( 16 , 8 , 4 ) & 4.117367040@xmath143 + ( 18 , 1 , 1 ) & @xmath233.093983511@xmath147 & ( 8 , 3 , 2 ) & @xmath237.739781828@xmath143 & ( 16 , 6 , 4 ) & @xmath233.212472559 & ( 18 , 8 , 4 ) & @xmath234.052076547@xmath143 + ( 4 , 2 , 2 ) & 2.370370370@xmath143 & ( 10 , 3 , 2 ) & @xmath231.053007756 & ( 18 , 6 , 4 ) & 4.903204193@xmath148 & ( 18 , 9 , 5 ) & 3.670192948@xmath143 + ( 6 , 2 , 2 ) & @xmath235.237389771@xmath143 & ( 12 , 3 , 2 ) & 1.967483048@xmath148 & ( 14 , 7 , 5 ) & 1.525642058@xmath143 & ( 10 , 5 , 0 ) & 3.248825624@xmath143 + ( 8 , 2 , 2 ) & 1.010441376 & ( 14 , 3 , 2 ) & @xmath231.305817944@xmath149 & ( 16 , 7 , 5 ) & @xmath234.034853659@xmath143 & ( 12 , 5 , 0 ) & @xmath236.899424873@xmath143 + ( 10 , 2 , 2 ) & @xmath231.494008026 & ( 16 , 3 , 2 ) & 6.744405152@xmath149 & ( 18 , 7 , 5 ) & @xmath232.770916581 & ( 14 , 5 , 0 ) & @xmath236.463536863 + ( 12 , 2 , 2 ) & 1.846909265@xmath143 & ( 18 , 3 , 2 ) & @xmath232.972249754@xmath150 & ( 16 , 8 , 6 ) & 1.176390583@xmath143 & ( 16 , 5 , 0 ) & 6.567477615@xmath148 + ( 14 , 2 , 2 ) & 1.042676508@xmath148 & ( 8 , 4 , 3 ) & 1.888174090@xmath143 & ( 18 , 8 , 6 ) & @xmath233.342928209@xmath143 & ( 18 , 5 , 0 ) & @xmath233.388841937@xmath149 + ( 16 , 2 , 2 ) & @xmath231.111625374@xmath149 & ( 10 , 4 , 3 ) & @xmath232.886375899@xmath143 & ( 18 , 9 , 7 ) & 9.175482371@xmath142 & ( 12 , 6 , 1 ) & 6.698993726@xmath143 + ( 18 , 2 , 2 ) & 1.249271047@xmath150 & ( 12 , 4 , 3 ) & @xmath231.577708121 & ( 6 , 3 , 0 ) & 3.732392710@xmath143 & ( 14 , 6 , 1 ) & @xmath231.079421253 + ( 6 , 3 , 3 ) & 9.330981775@xmath142 & ( 14 , 4 , 3 ) & 1.855760450@xmath148 & ( 8 , 3 , 0 ) & @xmath239.806190570@xmath143 & ( 16 , 6 , 1 ) & @xmath232.116718725@xmath148 + ( 8 , 3 , 3 ) & @xmath232.869718840@xmath143 & ( 16 , 4 , 3 ) & @xmath231.206715261@xmath149 & ( 10 , 3 , 0 ) & @xmath237.606578761@xmath143 & ( 18 , 6 , 1 ) & 2.016971553@xmath149 + ( 10 , 3 , 3 ) & 7.302649849@xmath143 & ( 18 , 4 , 3 ) & 6.228213110@xmath149 & ( 12 , 3 , 0 ) & 1.900360917@xmath148 & ( 14 , 7 , 2 ) & 6.538465963@xmath143 + ( 12 , 3 , 3 ) & 6.474454323@xmath142 & ( 10 , 5 , 4 ) & 1.082941875@xmath143 & ( 14 , 3 , 0 ) & @xmath231.096944607@xmath149 & ( 16 , 7 , 2 ) & @xmath236.381513372@xmath143 + ( 14 , 3 , 3 ) & @xmath231.507976852@xmath148 & ( 12 , 5 , 4 ) & @xmath237.182655253@xmath142 & ( 16 , 3 , 0 ) & 4.525126033@xmath149 & ( 18 , 7 , 2 ) & @xmath232.878600133@xmath148 + ( 16 , 3 , 3 ) & 1.112739551@xmath149 & ( 14 , 5 , 4 ) & @xmath231.662709009 & ( 18 , 3 , 0 ) & @xmath231.604127694@xmath150 & ( 16 , 8 , 3 ) & 6.176050560@xmath143 + ( 18 , 3 , 3 ) & @xmath235.501017323@xmath149 & ( 16 , 5 , 4 ) & 1.713732791@xmath148 & ( 8 , 4 , 1 ) & 5.664522270@xmath143 & ( 18 , 8 , 3 ) & @xmath231.660697208@xmath143 + ( 8 , 4 , 4 ) & 4.720435225@xmath142 & ( 18 , 5 , 4 ) & @xmath231.134996640@xmath149 & ( 10 , 4 , 1 ) & @xmath231.358637239 & ( 18 , 9 , 4 ) & 5.709189031@xmath143 + ( 10 , 4 , 4 ) & @xmath231.753576976@xmath143 & ( 12 , 6 , 5 ) & 6.698993726@xmath142 & ( 12 , 4 , 1 ) & @xmath235.476288758 & ( 12 , 6 , 0 ) & 3.349496863@xmath143 + ( 12 , 4 , 4 ) & 5.665451332@xmath143 & ( 14 , 6 , 5 ) & 2.348849413@xmath142 & ( 14 , 4 , 1 ) & 6.555818772@xmath148 & ( 14 , 6 , 0 ) & @xmath232.248529752@xmath143 + ( 14 , 4 , 4 ) & 7.031489323@xmath143 & ( 16 , 6 , 5 ) & @xmath231.498021733 & ( 16 , 4 , 1 ) & @xmath233.561148091@xmath149 & ( 16 , 6 , 0 ) & @xmath231.304144759@xmath148 + ( 16 , 4 , 4 ) & @xmath232.234394856@xmath148 & ( 18 , 6 , 5 ) & 1.514875202@xmath148 & ( 18 , 4 , 1 ) & 1.327749407@xmath150 & ( 18 , 6 , 0 ) & 1.069591051@xmath149 + ( 18 , 4 , 4 ) & 1.696295810@xmath149 & ( 14 , 7 , 6 ) & 4.358977309@xmath142 & ( 10 , 5 , 2 ) & 4.331767498@xmath143 & ( 14 , 7 , 1 ) & 7.628210290@xmath143 + ( 10 , 5 , 5 ) & 2.707354687@xmath142 & ( 16 , 7 , 6 ) & 6.397044909@xmath142 & ( 12 , 5 , 2 ) & @xmath238.287589771@xmath143 & ( 16 , 7 , 1 ) & @xmath231.648715187@xmath143 + ( 12 , 5 , 5 ) & @xmath231.145505796@xmath143 & ( 18 , 7 , 6 ) & @xmath231.267992012 & ( 14 , 5 , 2 ) & @xmath238.117281873 & ( 18 , 7 , 1 ) & @xmath233.918669310@xmath148 + ( 14 , 5 , 5 ) & 4.544498080@xmath143 & ( 16 , 8 , 7 ) & 2.940976457@xmath142 & ( 16 , 5 , 2 ) & 8.399862105@xmath148 & ( 16 , 8 , 2 ) & 8.234734079@xmath143 + ( 16 , 5 , 5 ) & 1.078497914 & ( 18 , 8 , 7 ) & 7.770975088@xmath142 & ( 18 , 5 , 2 ) & @xmath234.609565870@xmath149 & ( 18 , 8 , 2 ) & 1.073653528@xmath143 + ( 18 , 5 , 5 ) & @xmath232.724985485@xmath148 & ( 18 , 9 , 8) & 2.038996082@xmath142 & ( 12 , 6 , 3 ) & 3.349496863@xmath143 & ( 18 , 9 , 3 ) & 8.563783546@xmath143 + ( 12 , 6 , 6 ) & 1.674748432@xmath142 & ( 4 , 2 , 0 ) & 4.740740741@xmath143 & ( 14 , 6 , 3 ) & @xmath234.223690648@xmath143 & ( 14 , 7 , 0 ) & 4.358977309@xmath143 + ( 14 , 6 , 6 ) & @xmath237.829101098@xmath142 & ( 6 , 2 , 0 ) & @xmath236.512874780@xmath143 & ( 16 , 6 , 3 ) & @xmath239.671323639 & ( 16 , 7 , 0 ) & @xmath234.703472705@xmath143 + ( 16 , 6 , 6 ) & 3.741008913@xmath143 & ( 8 , 2 , 0 ) & @xmath231.556811959 & ( 18 , 6 , 3 ) & 9.548744493@xmath148 & ( 18 , 7 , 0 ) & @xmath231.979896930@xmath148 + ( 18 , 6 , 6 ) & 1.267008305 & ( 10 , 2 , 0 ) & 1.207544510@xmath148 & ( 14 , 7 , 4 ) & 2.615386385@xmath143 & ( 16 , 8 , 1 ) & 1.029341760 + ( 14 , 7 , 7 ) & 1.089744327@xmath142 & ( 12 , 2 , 0 ) & @xmath233.196277031@xmath148 & ( 16 , 7 , 4 ) & @xmath231.332433821@xmath143 & ( 18 , 8 , 1 ) & @xmath234.479185205@xmath143 + ( 16 , 7 , 7 ) & @xmath235.528545914@xmath142 & ( 14 , 2 , 0 ) & @xmath235.271684834@xmath148 & ( 18 , 7 , 4 ) & @xmath231.033273379@xmath148 & ( 18 , 9 , 2 ) & 1.141837806 + ( 18 , 7 , 7 ) & 3.138641071@xmath143 & ( 16 , 2 , 0 ) & 1.018254901@xmath150 & ( 16 , 8 , 5 ) & 2.058683520@xmath143 & ( 16 , 8 , 0 ) & 5.146708800@xmath143 + ( 16 , 8 , 8) & 7.352441142@xmath144 & ( 18 , 2 , 0 ) & @xmath235.459613629@xmath150 & ( 18 , 8 , 5 ) & 6.203359335@xmath142 & ( 18 , 8 , 0 ) & 2.583076501@xmath143 + ( 18 , 8 , 8) & @xmath234.001373177@xmath142 & ( 6 , 3 , 1 ) & 5.598589065@xmath143 & ( 18 , 9 , 6 ) & 1.631196866@xmath143 & ( 18 , 9 , 1 ) & 1.284567532 + ( 18 , 9 , 9 ) & 5.097490206@xmath144 & ( 8 , 3 , 1 ) & @xmath239.918360628@xmath143 & ( 8 , 4 , 0 ) & 2.832261135@xmath143 & ( 18 , 9 , 0 ) & 7.136486288@xmath143 + | we study one and two hole properties of the @xmath0 model at half - filling on the square lattice using series expansion methods at @xmath1 .
the dispersion curve for one hole excitations is calculated and found to be qualitatively similar to that obtained by other methods , but the bandwidth for small @xmath2 is some 20% larger than given previously .
we also obtain the binding energy and dispersion relation for two hole bound states .
the lowest bound state as @xmath2 increases is found to be first @xmath3-wave , and then @xmath4-wave , in accordance with predictions based upon the kohn - luttinger effect .
we also make a similar study for the @xmath5 model . |
application of particle induced x - ray emission ( pixe ) to non - destructive trace element analysis of materials has first been proposed by johansson and co - workers in 1970 @xcite .
today , this experimental technique is widely exploited in diverse fields @xcite
. the physical process of pixe may also give rise to unwanted instrumental background x - ray lines , as is the case for space missions and for some laboratory environments .
it also affects the spatial distribution of the energy deposit associated with the passage of charged particles in matter : in this respect , its effects may become significant in the domain of microdosimetry .
the wide application of this experimental technique has motivated the development of several dedicated software systems ; nevertheless , despite its large experimental interest , limited functionality for pixe simulation is available in general - purpose monte carlo codes .
this paper discusses the problem of simulating pixe in the context of a general - purpose monte carlo system and describes a set of developments to improve its simulation with geant4 @xcite .
finally , it illustrates an application of the developed pixe simulation prototype to the optimizationof the passive shield of the x - ray detectors of the erosita @xcite ( extended roentgen survey with an imaging telescope array ) telescope on the _ spectrum - x - gamma _ @xcite space mission .
software tools are available in support of pixe experimental applications as specialized codes or included in general - purpose simulation systems .
dedicated pixe codes are focussed on the application of this technique to elemental analysis .
they are concerned with the calculation of experimentally relevant x - ray yields resulting from the irradiation of a material sample by an ion beam : primarily transitions concerning the k shell , and in second instance transitions originating from vacancies in the l shell . for this purpose various analysis programs
have been developed , which are able to solve the inverse problem of determining the composition of the sample from an iterative fitting of a pixe spectrum ; some among them are geopixe @xcite , gupix @xcite , pixan @xcite , pixeklm @xcite , sapix @xcite , winaxil @xcite and wits - hex @xcite .
a few codes concern pixe simulation @xcite specifically .
these codes share basic physics modelling options , like the adoption of the ecpssr ( energy - loss coulomb - repulsion perturbed stationary state relativistic ) @xcite model for the calculation of ionization cross sections ; they handle simple experimental geometries , such as target materials consisting of layers , and impose limitations on the type of samples that can be analyzed .
dedicated pixe software systems have a limited application scope , as they lack the capability of dealing with complex experimental configurations .
comprehensive modelling capabilities are usually associated with general - purpose monte carlo systems .
however , the simulation of pixe is not widely covered by large scale monte carlo codes treating hadron interactions , while the conceptually similar simulation of electron impact ionisation is implemented in the egs @xcite and penelope @xcite electron - photon monte carlo systems .
the geant4 simulation toolkit addresses x - ray emission induced both by electrons and heavy particles like protons and @xmath0 particles .
the physics that needs to be considered for the simulation of pixe includes the energy loss and scattering of the incident charged particle , atomic shell ionization cross sections , and atomic transition probabilities and energies . on top of these physics features ,
consistency should be ensured , when modelling pixe , with the particle transport schemes governing the monte carlo simulation .
intrinsically , pixe is a discrete process : x - ray emission occurs as the result of producing a vacancy in atomic shell occupancy , in competition with auger electron emission and coster - kronig transitions .
nevertheless , this discrete process is intertwined with the ionization process , which determines the production of the vacancy ; this process , for reasons which are elucidated below , is treated in general - purpose monte carlo codes with mixed condensed and discrete transport schemes .
the transport scheme to which ionization is subject affects the simulation of pixe .
the simulation of the energy loss of a charged particle due to ionization is affected by the infrared divergence of the cross section for producing secondary electrons . in the context of general purpose monte carlo systems ,
a discrete treatment of each individual ionization process becomes inhibitive , since , due to their large number , the required computation time becomes excessive .
infrared divergence is usually handled in general - purpose monte carlo codes by adopting a condensed - random - walk scheme @xcite for particle transport .
in such a scheme , the particle s energy loss and deflection are treated as averaged net effect of many discrete interactions along the step , thereby substituting in the simulation a single continuous process for the many discrete processes that actually occur . in a mixed scheme , like the one adopted by geant4 @xcite ,
two different rgimes of particle transport are introduced , which are distinguished through a secondary production threshold setting , i.e. a threshold for the kinetic energy of the electron that is kicked out of an atom as a result of ionization ( the so - called @xmath1-ray ) : all ionizations that would generate @xmath1-rays below the threshold are treated as a continuous process along the step , while the ionizations that produce @xmath1-rays above the threshold are treated as a discrete process .
while this combined condensed - random - walk and discrete particle transport scheme is conceptually appealing and appropriate to many simulation applications , it suffers from drawbacks with respect to the generation of pixe .
one drawback is that atomic relaxation occurs only in connection with the discrete part of the transport scheme , where the event of producing a @xmath1-ray can be associated with the creation of a vacancy in the shell occupancy .
for the same reason , the fluorescence yield depends on the threshold for the kinetic energy of the secondary electron .
another drawback of the current transport scheme is that the cross section for discrete hadron ionization , i.e. for production of a @xmath1-ray , is calculated from a model for energy loss that is independent of the shell where the ionization occurs . while theoretical calculations are available to determine the spectrum of the emitted electron for each sub - shell in the case of electron impact ionization @xcite for any element , to the authors knowledge no such facilities are currently available for the ionization induced by protons and ions .
experimental data are not adequate either to complement the lack of theoretical calculations .
at the present time , the geant4 toolkit does not provide adequate capabilities for the simulation of pixe in realistic experimental use cases .
the first development cycle @xcite had a limited scope : the implementations concerned only pixe induced by @xmath0 particles and involving k shell ionization - apart from the implementation based on gryzinski s @xcite theoretical model , which produces physically incorrect results .
even with the models which calculate k shell ionization cross sections correctly , inconsistencies arise in the simulation of pixe related to the algorithm implemented for determining the production of a k shell vacancy .
the deficiencies exhibited by the software released in geant4 9.2 @xcite did not contribute to improve pixe simulation with respect to the previous version .
a set of developments for pixe simulation is described in the following sections .
a preliminary overview of their progress was reported in @xcite .
the physics aspects associated with pixe involve the creation of a vacancy in the shell occupancy due to ionization , and the emission of x - rays from the following atomic deexcitation .
the former requires the knowledge of ionization cross sections detailed at the level of individual shells : for this purpose several models have been implemented and validated against experimental data .
the latter exploits the existing functionality of geant4 atomic relaxation package @xcite .
the domain decomposition at the basis of pixe simulation with geant4 identified three main entities with associated responsibilities : the hadron ionization process , the creation of a vacancy in the shell occupancy resulting from ionisation , the deexcitation of the ionised atom with the associated generation of x - rays .
the simulation of pixe is the result of the collaboration of these entities .
a class diagram in the unified modelling language ( uml ) @xcite illustrates the main features of the software design in fig .
[ uml_fig ] . the simulation of pixe concerns a variety of experimental applications , that require the capability of calculating ionisation cross sections over an extended energy range : from a few mev typical of material analysis applications to hundreds mev or gev range of astrophysical applications .
various theoretical and empirical models are available in literature to describe ionisation cross sections for different interacting particles , as well as compilations of experimental data . however , there is limited documentation in literature of systematic , quantitative assessments of the accuracy of the various models .
the current software prototype has adopted the strategy of providing an extensive collection of ionisation cross section models as a function of element , atomic ( sub-)shell , and incident particle kinetic energy . according to the chosen strategy ,
the provision of a cross section model is reduced to the construction of tabulations of its values at preset energies .
the cross sections associated with the models described in this paper have been pre - calculated either using existing software documented in literature , or developing ad hoc code . the data are stored in files , which make up a data library for pixe simulation with geant4 - but could be used also by other codes .
the cross section data sets selected by the user are loaded into memory at runtime ; cross section values at a given energy are calculated by interpolation over the tabulated values whenever required .
the adopted data - driven approach for the provision of ionization cross sections presents various advantages .
it optimizes performance speed , since the calculation of the interpolation is faster than the calculation from complex algorithms implementing theoretical models .
this approach also offers flexibility : chosing a cross section model simply amounts to reading the corresponding set of data files ; adding a new set of cross sections simply amounts to providing the corresponding set of data files , which are handled transparently .
finally , the cross section data are transparent to the user : the files are accessible to the user and human readable . a wide choice of cross section models for k , l and m shell ionization is provided in the prototype software for protons and @xmath0 particles .
the availability of ionization cross section calculations and experimental data for outer shells is very limited in literature .
theoretical cross section models include plane wave born approximation ( pwba ) and variants of the ecpssr model : the original ecpssr formulation @xcite , ecpssr with united atom correction ( ecpssr - ua ) @xcite , ecpssr with corrections for the dirac - hartree - slater nature of the k shell @xcite ( ecpssr - hs ) , as well as calculations based on recent improvements to k shell cross section specific to high energy @xcite ( ecpssr - he ) .
the cross sections have been tabulated and assembled in a data library ; the values at a given energy are calculated by interpolation .
the tabulations corresponding to theoretical calculations span the energy range between 10 kev and 10 gev ; empirical models are tabulated consistently with their energy range of validity .
the adopted data - driven approach optimizes performance speed and offers flexibility for chosing a cross section model .
ecpssr tabulations have been produced using the isics software @xcite , 2006 version and an extended version @xcite including recent high energy developments .
tabulations of ecpssr calculations as reported in @xcite are also provided .
empirical cross section models for k shell ionization include the tabulations for protons documented in @xcite and a more recent one @xcite . an empirical cross section model for k shell ionization by @xmath0 particles
is based on the tabulations in @xcite .
empirical models for l shell ionization by protons have been developed by miyagawa et al .
@xcite , sow et al .
@xcite and orlic et al .
@xcite .
the isics software allows the calculation of cross sections for heavier ions as well ; therefore , the current pixe simulation capabilities can be easily extended in future development cycles .
the determination of which atomic ( sub-)shell is ionised is related to its ionisation cross section with respect to the total cross section for ionising the target atom . however
, as previously discussed , the condensed - random - walk scheme raises an issue as to estimating the total ionisation cross section at a given energy of the incident particle .
a different algorithm has been adopted with respect to the one implemented in the first development cycle : the vacancy in the shell occupancy is determined based on the total cross section calculated by summing all the individual shell ionisation cross sections .
this algorithm provides a correct distribution of the produced vacancies as long as ionisation cross sections can be calculated for all the atomic shells involved in the atomic structure of the target element . since
cross section models are currently available for k , l and m shells only , at the present status of the software this algorithm overestimates pixe for elements whose atomic structure involves outer shells , because of the implicit underestimation of the total ionization cross section .
this approach , however , provides better control on the simulation results than the algorithm implemented in the first development cycle .
the production of secondary particles by the atomic relaxation of an ionized atom is delegated to the atomic relaxation component .
the availability of a wide variety of cross section models for the first time in the same computational environment allowed a detailed comparative assessment of their features against experimental data .
the comparison of cross sections as a function of energy was performed for each element by means of the @xmath2 test .
contingency tables were built on the basis of the outcome of the @xmath2 test to determine the equivalent , or different behavior of model categories .
the input to contingency tables derived from the results of the @xmath2 test : they were classified respectively as `` pass '' or `` fail '' according to whether the corresponding p - value was consistent with a 95% confidence level .
the contingency tables were analyzed with fisher exact test @xcite .
the reference experimental data were extracted from @xcite .
an example of how the simulation models reproduce experimental measurements is shown in fig .
[ fig_kexp ] . & kahoul + + 95% & 67 & 74 & 77 & 68 & 71 & 46 + 99% & 85 & 83 & 83 & 85 & 80 & 57 + + 95% & 69 & 75 & 86 & 69 & 70 & 48 + 99% & 83 & 81 & 91 & 83 & 80 & 56 + the fraction of test cases for which the @xmath2 test fails to reject the null hypothesis at the 95% and 99% confidence level are listed in table [ tab_kpass ] : all the cross section models implemented in the simulation exhibit equivalent behaviour regards the compatibility with the experimental data , with the exception of the kahoul et al
the contingency table comparing the kahoul et al .
and ecpssr - hs models confirms that the two models show a statistically significant difference regards their accuracy ( p - value of 0.001 ) .
the contingency tables associated with the other models show that they are statistically equivalent regarding their accuracy .
however , when only the lower energy range ( below 5 - 7 mev , depending on the atomic number ) is considered , a statistically significant difference at the 95% confidence level ( p - value of 0.034 ) is observed between the ecpssr model and the ecpssr - hs one ; the latter is more accurate with respect to experimental data . from this analysis
one can conclude that the implemented k shell ionization cross section models exhibit a satisfactory accuracy with respect to experimental measurements .
the cross sections for l sub - shell ionization cross sections were compared to the experimental data collected in two complementary compilations @xcite , @xcite . an example of how the simulation models reproduce experimental measurements
is shown in fig .
[ fig_lexp ] .
the same method was applied as described for the validation of k shell cross sections .
the ecpssr model appears to provide a satisfactory representation of l shell ionisation cross sections with respect to experimental data , especially with its united atom variant .
the ecpssr - ua exhibits the best overall accuracy among the various models ; the orlic et al .
model exhibits the worst accuracy with respect to experimental data .
this semi - empirical model is the only option implemented in geant4 9.2 for the calculation of l shell ionization cross sections .
the accuracy of the various cross section models was studied by means of contingency tables to evaluate their differences quantitatively .
the categorical analysis was performed between the ecpssr model with united atom correction , i.e. the model showing the best accuracy according to the results of the @xmath2 test , and the other cross section models .
the contingency tables were built based on the results of the @xmath2 test at the 95% confidence level , summing the `` pass '' and `` fail '' outcome over the three sub - shells .
the orlic et al .
semi - empirical model is found to be significantly less accurate than the ecpssr - ua model : the hypothesis of equivalence of their accuracy with respect to experimental data is rejected at 99% confidence level .
the p - values concerning the comparison of the miyagawa et al . empirical model
are close to the critical region for 95% confidence , and slightly different for the three tests performed on the related contingency table .
the sow et al . empirical model and the ecpssr model in its original
formulation appears statistically equivalent in accuracy to the ecpssr model with united atom correction . as a result of this analysis , the ecpssr model with united atom approximation
can be recommended for usage in geant4-based simulation applications as the most accurate option for l shell ionization cross sections .
the ecpssr model in its original formulation can be considered a satisfactory alternative ; the sow et al .
empirical model has satisfactory accuracy , but limited applicability regards the target elements and proton energies it can handle .
pixe as a technique for elemental analysis is usually performed with proton beams of a few mev . in the recent years
, higher energy proton beams of a few tens mev have been effectively exploited too .
hhigh energy protons are a source of pixe in the space radiation environment .
the interest in high energy pixe has motivated recent theoretical investigations @xcite concerning cross section calculations at higher energies . despite the emerging interest of high energy pixe ,
only a limited set of experimental data is available above the energy range of conventional pixe techniques .
the accuracy of the implemented k shell cross section models was evaluated against two sets of measurements at higher energy @xcite , respectively at 66 and 68 mev . the experimental measurement with uranium was not included in the comparison , since it appears affected by some experimental systematics .
the @xmath2 test was performed first separately on either experimental data set to evaluate the possible presence of any systematics in the two test cases , then on the combined data set .
the p - values from the @xmath2 test against these experimental data are listed in table [ tab_hepixe ] . over the limited data sample considered in this test
, the ecpssr model with the correction in @xcitethis model does not appear to provide better accuracy than the original ecpssr formulation ; nevertheless more high energy experimental data would be required to reach a firm conclusion .
also , this analysis should be verified over tabulation deriving from a published version of the isics code , when it becomes available . the @xmath2 test over the experimental data at 160 mev collected in @xcite results in p - values less than 0.001 for all the ecpssr model variants .
the rejection of the null hypothesis could be ascribed either to the deficiency of the theory or to systematic effects affecting the measurements ; further data would be required for a sound assessment . .p - values from the @xmath2 test concerning high energy experimental data [ cols="<,^,^,^,^ " , ]
the prototype components for pixe simulation described in the previous sections were applied to a study of the passive , graded z shielding of the x - ray detectors of the erosita telescope @xcite on the upcoming russian _
spectrum - x - gamma _ space mission .
the purpose of the passive shielding is two - fold .
firstly , the passive shielding prevents abundant low - energy cosmic - ray particles from reaching the detectors and thus from causing radiation damage .
secondly , the passive graded z shielding serves to reduce instrumental background noise to a minimum @xcite .
this background noise consists of both fluorescence lines and continuum background due to bremsstrahlung photons and @xmath1 rays from cosmic - ray interactions .
the event timing accuracy of current imaging si detectors for x - ray astronomy ( photon energy range @xmath3 kev ) is limited by the signal integration time of these devices . for such detectors ,
an active anti - coincidence system can not be used for background reduction because discarding complete readout frames correlated with an anti - coincidence signal would result in unacceptable dead time .
detector triggers due to primary cosmic - ray particles can be discriminated in imaging detectors due to their high energy deposit and their pixel image pattern .
however , interactions of primary cosmic - ray particles in the detector and surrounding passive material give rise to secondary x - rays and charged particles .
these may in turn lead to detector triggers within the accepted energy range .
such triggers contribute to the instrumental background noise because they can not be distinguished from valid events due to cosmic x - ray photons that were focused by the telescope mirror system onto the detector .
the production of secondary photons and particles by the ubiquitous cosmic radiation is inevitable , but graded z shielding permits the shifting of the energy of secondaries from atomic deexcitation to low values .
the probability that an atomic shell vacancy is filled by radiative transitions ( emission of fluorescence x - ray photons ) decreases with decreasing atomic charge number z ; by contrast , the probability for non - radiative transitions ( emission of auger and coster - kronig electrons ) increases .
the average energy of secondaries from atomic deexcitation decreases with decreasing charge number z. therefore , in a graded z shield cosmic - ray induced fluorescence x rays produced in an outer , higher z layer of the shield are absorbed in an inner , lower z layer .
subsequent atomic deexcitation in this inner layer gives rise to fluorescence photons and auger electrons with energies that are lower than the energies of the deexcitation particles from the outer layer ; in addition , there will be relatively more deexcitation electrons than x - rays .
if the effective charge number z of the innermost shield layer is sufficiently low , ionization results in the generation of mainly auger electrons with energies below 1 kev , which can easily be stopped in a thin passivation layer on top of the detectors . ionization can also create rare fluorescence x rays of similarly low energy .
a first set of graded z shield designs was studied by monte carlo simulation , using the pixe prototype software together with geant4 versions 9.1-patch 01 .
the detector chip was modelled as a 450 @xmath4 m thick square slab of pure si with dimensions 5.6 cm @xmath5 3.7 cm
. the sensitive detector , which is integrated into the chip , covers an area of 2.88 cm @xmath5 2.88 cm or 384 @xmath5 384 square pixels of size 75 @xmath4 m @xmath575 @xmath4 m .
this detector model was placed inside a box - shaped shield ; figures of the actual design can be found in @xcite . in its simplest form ,
the passive shield consisted only of a single 3 cm thick layer of pure cu ; the outer dimensions of the shield were 12.7 cm @xmath5 10.8 cm @xmath5 7.1 cm .
a second version of the graded z shield had a 1 mm thick al layer inside the cu layer , and a third version in addition a 1 mm thick b@xmath6c layer inside the al layer .
the physics configuration in the simulation application involved the library - based processes of the geant4 low energy electromagnetic package for electrons and photons , along with the improved version of the hadron ionisation process and the specific pixe software described in section [ sec_newpixe ] . among the ionization cross section models ,
the ecpssr ones were selected for k , l and m shells .
_ spectrum - x - gamma _ is expected be launched in 2012 into an l2 orbit .
the background spectra due to cosmic - ray protons were simulated for the three different erosita graded z shield designs . a model for the detector response , taking into account fano statistics ( and hence the energy resolution ) and detector noise ,
was then applied to obtain a simulated data sets .
these simulated data were further processed with specialized data analysis algorithms @xcite .
a comparison of the results is depicted in fig .
[ comp_shield ] .
the spectra represent the average background in a detector pixel .
qualitatively , the pixe prototype implementation is working properly : protons produce the expected fluorescence lines with correct relative strengths . in case of a pure cu shield , shown in fig .
[ comp_cu - shield ] , strong cu k@xmath7 and k@xmath8 fluorescence lines at about 8.0 and 8.9 kev are present in the background spectrum . in case of a combined cu - al shield , depicted in fig .
[ comp_cu - al - shield ] , the cu fluorescence lines are absorbed in al , but ionization in al gives rise to a clear al k@xmath7 fluorescence line at about 1.5 kev . in case of the full graded z shield configuration , shown in fig .
[ comp_cu - al - b4c - shield ] , the b@xmath6c layer absorbs the al line , but at the same is not a source of significant fluorescence lines , which is expected due to the low fluorescence yield of these light elements .
except for an excess below 0.3 kev for the case of the cu - al - b@xmath6c graded z shield , which appears because the simulated detector model does not yet include a thin passivation layer , there is no significant difference in the continuum background for the three different designs .
the inclusion of a thin layer for the treatment of the low energy background will be the object of a further detector design optimization .
this application demonstrates that the developed software is capable of supporting concrete experimental studies .
nevertheless , the concerns outlined in sections [ sec_transport ] and [ sec_vacancy ] should be kept in mind : while the present pixe simulation component can provide valuable information in terms of relative fluorescence yields from inner shells , the intrinsic limitations of the mixed transport scheme in which ionization is modelled and the lack of cross section calculations for outer shells prevent an analysis of the simulation outcome in absolute terms . .
this paper presents a brief overview of the status , open issues and recent developments of pixe simulation with geant4 ; a more extensive report of the underlying concepts , developments and results is available in @xcite .
these new developments represent a significant step forward regards pixe simulation with geant4 .
they extend the capabilities of the toolkit by enabling the generation of pixe associated with k , l and m shells for protons and @xmath0 particles ; for this purpose a variety of cross section models are provided . the adopted data - driven strategy and the software design
improve the computational performance over previous geant4 models .
the validity of the implemented models has been quantitatively estimated with respect to experimental data .
an extensive ionisation cross section data library has been created as a by - product of the development process : it can be of interest to the experimental community for a variety of applications , not necessarily limited to pixe simulation with geant4 .
some issues identified in the course of the development process are still open : they concern the consistency of pixe simulation in a mixed condensed - discrete particle transport scheme . in parallel ,
a project @xcite is in progress to address design issues concerning co - working condensed and discrete transport methods in a general purpose simulation system . despite the known limitations related to mixed transport schemes , the software developments described in this paper provide sufficient functionality for realistic experimental investigations
the authors express their gratitude to a. zucchiatti for valuable discussions on pixe experimental techniques and to s. cipolla for providing a prototype version of isics including updates for high energy k shell cross sections .
the authors are grateful to the rsicc staff , in particular b. l. kirk and j. b. manneschmidt , for the support to assemble a ionization cross section data library resulting from the developments described in this paper .
j. baro et al . , `` penelope : an algorithm for monte carlo simulation of the penetration and energy loss of electrons and positrons in matter '' _ nucl .
b _ , vol .
31 - 46 , 2005 .
m. j. berger , `` monte carlo calculation of the penetration and diffusion of fast charged particles '' , in _ methods in computational physics _ , vol .
1 , ed . b. alder , s. fernbach and m. rotenberg , academic press , new york , pp .
135 - 215 , 1963 . s. t. perkins , d. e. cullen , and s. m. seltzer , `` tables and graphs of electron - interaction cross sections from 10 ev to 100 gev derived from the llnl evaluated electron data library ( eedl ) '' , ucrl-50400 vol . 31 , 1997 . s. j. cipolla , `` the united atom approximation option in the isics program to calculate k , l , and m - shell cross sections from pwba and ecpssr theory '' , _ nucl .
b _ , vol . 261 , pp . 142 - 144 , 2007 . s. j. cipolla , `` an improved version of isics : a program for calculating k , l , and m - shell cross sections from pwba and ecpssr theory using a personal computer '' , _ comp .
157 - 159 , 2007 .
y. miyagawa et al .
, `` analytical formulas for ionization cross sections and coster - kronig corrected fluorescence yields of the ll , l2 , and l3 subshells '' , _ nucl .
b _ , vol .
115 - 122 , 1988 .
n. meidinger et al .
, `` erosita camera design and first performance measurements with ccds '' , in `` space telescopes and instrumentation 2008 : ultraviolet to gamma ray '' , proc .
spie , vol .
7011 , 70110j , 2008 . | particle induced x - ray emission ( pixe ) is an important physical effect that is not yet adequately modelled in geant4 .
this paper provides a critical analysis of the problem domain associated with pixe simulation and describes a set of software developments to improve pixe simulation with geant4 .
the capabilities of the developed software prototype are illustrated and applied to a study of the passive shielding of the x - ray detectors of the german erosita telescope on the upcoming russian _
spectrum - x - gamma _ space mission .
monte carlo , geant4 , pixe , ionization . |
it is nowadays well recognized that reliable nuclear physics inputs are essential for realistic simulations of many astrophysical phenomena . depending on the particular time scales and the resulting equilibrium conditions
, this can be an equation of state ( eos ) and individual ( nuclear ) reactions , respectively . due to the relatively long timescale of reactions mediated by the weak interaction , the latter play an important role in many sites , for instance the late stages of massive star evolution @xcite , thermonuclear @xcite and core - collapse supernovae @xcite , nucleosynthesis and energy generation in x - ray bursts and other rp - process sites @xcite , the accreting neutron star crust @xcite , and neutron star mergers @xcite .
within this paper , we will concentrate on core - collapse supernovae ( ccsne ) .
except for very low - density matter encountered in the outer layers , time scales are such that strong and electromagnetic interactions are in equilibrium .
therefore , at any time , the composition of matter can be calculated as a function of the local temperature ( @xmath3 ) , baryon number density ( @xmath4 ) and proton fraction ( @xmath5 ) assuming nuclear statistical equilibrium ( nse ) ( see e.g. @xcite ) . on the contrary ,
weak interactions can in general not be considered in equilibrium and individual reactions rates are crucial to determine the local proton fraction .
in particular , electron capture determines @xmath5 in the first stages of the collapse , the associated size of the homologous core , and has thus an impact on the consequent explosion dynamics @xcite . for these simulations ,
the time and space dependent electron capture rates are obtained for a given @xmath6 by folding the nse nuclear distribution with the capture rates on individual nuclei .
the microphysics uncertainties on the rates thus originate both from the uncertainties on the nse distribution , and on those associated with the individual rates .
a very complete study on the sensitivity of core collapse dynamics to variations of electron capture rates on medium - heavy nuclei has recently been performed @xcite .
the authors thereby concentrate on the latter aspect , modifying the rates on individual nuclei according to present uncertainties , using a comprehensive set of progenitors and eos .
it was shown that important variations in the mass of the inner core at bounce and in the peak neutrino luminosity have to be expected .
the variations induced by the modified electron capture rates are five times larger than those induced by the uncertainty on the progenitor model , showing the importance of an increased reliability of nuclear physics inputs . in this same work
it was clearly shown that the results are most sensitive to the region around the @xmath0 shell closure for very neutron rich nuclei ( @xmath7 ) . in this work
we address the complementary aspect of uncertainties associated with the nuclear distribution .
specifically , nuclear magicity , as incorporated in mass models within currently available eos , is known to be deeply modified in very neutron rich nuclei @xcite and the pronounced shell closures at @xmath0 and @xmath8 are expected to be quenched far from stability .
we study here the impact of such a possible quenching on the matter composition during collapse , and the associated modification of the electron capture rate .
we show that modifications of the electron capture rate of up to 30% are possible in the case of strong shell quenching . unlike ref .
@xcite , in this exploratory calculation we do not aim at computing the complete time evolution of @xmath5 , nor the associated modification of the inner core mass and the neutrino luminosity . such a complete simulation would require a consistent model for the mass and electron capture rates as determined by fermi and gamow - teller transition strengths for the relevant nuclei .
this is beyond the scope of the present paper .
however , from general arguments and the very detailed results of ref .
@xcite , we expect that an increased capture rate leads to enhanced neutrino cooling , an accelerated collapse , and a higher inner core mass at bounce .
the main message of the present work is to stress the need of structure information and mass measurements , particularly around the @xmath2ni region . from the theory viewpoint
, it appears very important to have microscopic and consistent calculations of both masses and weak rates .
the paper is organized as follows . in section
[ sec : sec2 ] the different microphysical inputs entering the electron capture rate calculations are presented , namely the individual rates and the equation of state .
we show that for the latter , the mass model represents the key ingredient .
the influence of the mass model on the capture rates during core collapse is discussed in section [ sec : sec3 ] . to that end , we consider two representative collapse trajectories with typical thermodynamic conditions .
an effective parametrization simulating the possible shell quenching far from stability of the two relevant shell closures @xmath0 and @xmath1 is introduced , and the associated modification of the global rate is discussed .
section [ sec : sec4 ] contains summary and conclusions . throughout the paper we will use units such that @xmath9
as mentioned above , for given thermodynamic condition , i.e. given values of @xmath6 , the total electron capture rate is obtained by folding individual rates with the nuclear distribution .
the latter is obtained by nse calculations which depend in turn on a number of inputs and in particular on the masses of different nuclei , thus on the mass model employed . in this section
we detail the different ingredients employed in this work .
we will start with the individual rates in sec [ sec : individual ] .
the nse model will be discussed in sec .
[ sec : nse ] .
[ sec : mass ] will be devoted to the mass model and special attention will be paid to the possible quenching of magicity in very neutron rich nuclei .
concerning the individual rates , tabulated values are available from large scale shell model calculations in the @xmath10-shell @xcite ( ohmts ) , and @xmath11-shell @xcite ( lmp ) .
since those calculations are still nowadays numerically very demanding , for heavier nuclei , the shell model rates are complemented with shell model monte carlo and rpa calculations @xcite or an empirical approach @xcite ( pf ) . in ref .
@xcite shell model rates on additional nuclei have been calculated , however , they are unfortunately not available as individual rates .
these extended calculations still comprise a limited range of elements , masses , temperature and electron densities ( @xmath12 ) and are not sufficient to cover completely the typical conditions in the most central part of the core collapse , nor the typical nuclei encountered in those situations .
we have therefore decided to use here the analytical equation proposed by langanke et al .
( l03 ) in ref .
it is based on a parametrization of the electron - capture rate as a function of the ground state to ground state @xmath13-value and writes , @xmath14 , \label{eq : l03}\end{aligned}\ ] ] with @xmath15 s , @xmath16 , @xmath17 .
@xmath3 represents the temperature , @xmath18 and @xmath19 stand for electron rest mass and , respectively , chemical potential .
@xmath20 denotes the relativistic fermi integral , @xmath21 .
@xmath22 stands for a typical ( gamow - teller plus forbidden ) matrix element and @xmath23 represents the energy difference between the parent state @xmath24 and the daughter state @xmath25 .
their values , obtained by fitting microscopic data , are @xmath26 mev and @xmath27 mev . for electron
capture rates on protons we use the results of ref .
@xcite .
the first approximation of weak interaction rates in terms lepton chemical potential and @xmath13-value was proposed by fuller , fowler and newman in the pioneering ref .
@xcite in the form @xmath28 , where @xmath29 is the space factor .
nuclear structure effects were there approximately accounted for via the effective log(ft)-value .
for ec - rates three values have been proposed based on the neutron and proton numbers of the parent nucleus ( upon which nuclei are classified in `` blocked '' ( @xmath30 or @xmath31 ) and `` unblocked '' ( @xmath32 and @xmath33 ) ) and the relation between electron chemical potential and @xmath13-value . for `` unblocked '' nuclei ec was supposed to occur mainly through the gamow - teller resonances which are roughly at an excitation of 3 mev above the daughter ground state . )
( a , d ) , lmp @xcite ( @xmath34 ) ( b , e ) and pf @xcite ( @xmath35 ) ( c , f ) tables ( solid red circles ) for @xmath3=0.68 mev , @xmath36 @xmath37 , @xmath38=0.447 ( a , b , c ) and @xmath3=1.30 mev , @xmath39 @xmath37 , @xmath38=0.361 ( d , e , f ) . ]
the validity of the analytic equation l03 has been recurrently addressed in the literature ( see e.g. @xcite ) and proved to offer a fair approximation under thermodynamical conditions relevant for core - collapse .
scatter of microscopic rates around the @xmath13-dependence at low values of temperature and electron density were interpreted in ref . @xcite as indicating that several parent and daughter states with different transition strengths contribute to the process . on the contrary ,
better agreement at higher values of temperature and electron density was attributed to a value of the electron chemical potential high enough to make ec - rate independent of the strength distribution .
[ fig : ec_table_l03 ] addresses the same issue by confronting the predictions of eq .
( [ eq : l03 ] ) , plotted as a function of reaction heat @xmath13 , with those obtained by linear interpolation of more microscopic and sophisticated weak interaction rate tables ohmts @xcite ( panels ( a ) and ( d ) ) , lmp @xcite ( panels ( b ) and ( e ) ) and pf @xcite ( panels ( c ) and ( f ) ) .
two representative thermodynamic conditions taken from the core - collapse trajectory of a @xmath40 star in ref .
@xcite are considered ( see figure caption ) . as already evidenced by other authors , fig .
[ fig : ec_table_l03 ] shows a correct overall behavior of @xmath41 and , in what concerns the individual rates , important deviations .
quite remarkably , eq.([eq : l03 ] ) offers a fair approximation also outside the mass range for which it was designed . whenever exist the deviations underline the well known fact that , in order to accurately describe astrophysical processes , it will be very important to have fully microscopic calculations covering the whole mass , charge , temperature and electron fraction domain .
it was recently demonstrated in ref .
@xcite that a global arbitrary re - scaling of a factor 2,3 or 10 of the unknown rates leads to important modifications in the dynamics of the collapse .
given the purpose of this paper , which is to analyze the influence of the mass extrapolations in the still experimentally unknown region of very neutron rich nuclei , we have not tried to play with these rates , and we consider that for our purpose the l03 equations are sufficiently accurate to provide a reference calculation .
the model is based on a statistical distribution of compressible nuclear clusters composed by @xmath42 nucleons ( @xmath43 neutrons and @xmath44 protons ) immersed in a homogeneous background of self - interacting nucleons and electrons .
the details of the model are explained elsewhere @xcite ; here we only recall the main physical ingredients which are important for the present study .
the different thermodynamic quantities in the baryonic sector are decomposed into the sum of a term pertaining to the nucleon gas , and a term given by the contributions of the distribution of nuclei .
let us start with the gas contribution . in absence of clusters ,
it would simply be given by the free energy density of homogeneous nuclear matter at density @xmath45 , asymmetry @xmath46 and temperature @xmath3 .
in the non - relativistic mean field approximation it reads ( @xmath47 ) : @xmath48 with the entropy density given by : @xmath49 \label{eq : s_hm}\end{aligned}\ ] ] in these equations , @xmath50 is the spin degeneracy in spin - saturated matter , @xmath51 is the finite temperature occupation number at effective chemical potential @xmath52 , @xmath53^{-1}$ ] . the choice for the energy density @xmath54 and the effective nucleon mass @xmath55 , containing the interaction effects , defines the equation of state . for the applications shown in this paper
we have considered a large set of different well - known skyrme functionals for these two quantities @xcite , which have been successfully confronted with different nuclear structure data and are also compatible with the most recent experimental constraints on nuclear matter properties @xcite . in the present nse model , the free energy density of the nucleon gas
is reduced with respect to the homogeneous gas expression , eq .
( [ eq : f_hm ] ) , in order to account for the in - medium modification due to the finite volume occupied by the clusters @xcite .
the result is : @xmath56 where the sum runs over the different clusters weighted by their multiplicity per unit volume @xmath57 .
@xmath58 denotes the saturation density of asymmetric matter evaluated at the cluster asymmetry @xmath59 , and accounts for the compressible character of the clusters .
the asymmetry @xmath59 differs from the cluster global asymmetry @xmath60 because of coulomb and skin effects , and is additionally modified to account for the influence of the external gas @xcite .
we have checked that , within these phenomenological limits , the choice of the interaction does not produce any sensible effect neither on the composition of matter during collapse nor on the average capture rates .
this is due to the fact that , during the collapse trajectories we have considered , the density of the nucleonic gas and its neutron enrichment are not important enough to make unbound nucleon energetics play a significant role .
this result is perfectly compatible with previous findings @xcite .
the cluster multiplicities are given by the self - consistent nse expression : @xmath61 where we have defined the bound fraction of clusters by @xmath62 , @xmath63 , with @xmath64 .
the chemical potentials can be expressed as a function of the gas densities only : @xmath65 in eq.([mult_nse ] ) @xmath66 is the free energy of the cluster immersed in the nucleon gas : @xmath67 where the total volume @xmath68 has been introduced .
again , because of the low gas densities considered in this application , the third and fifth term in eq .
( [ fenergy_cl_ws ] ) , representing the in - medium bulk and surface modification of the cluster free energy due to the presence of the gas , turn out to be negligible .
the most important ingredient of the nse model is thus given by the vacuum binding energy of the cluster @xmath22 , corrected by the well known electron screening effect @xmath69 , and , to less extent , the cluster level density entering the temperature dependent degeneracy factor @xmath70 given by : @xmath71 , \label{deg_bucurescu}\ ] ] where @xmath72 is the density of states of the cluster , @xmath73 is the average particle separation energy , @xmath74 is the nucleon mass .
concerning the density of states , we use a back - shifted fermi gas model with parameters fitted from experimental data @xcite .
the key quantity of the model is thus the cluster binding energy , which we shall discuss in more detail in the following subsection .
consistency with the equation of state of the free nucleons in principle demands that the cluster mass should be evaluated with the same energy functional employed for the gas .
this is indeed done in the most recent nse models @xcite .
however , no functional model sufficiently precisely reproduces nuclear ground states , and when experimentally measured nuclear masses are available , they are preferentially used in all recent nse models @xcite .
it is also important to stress at this point that , independent of the nuclear matter parameters of the eos ( symmetric nuclear matter incompressibility @xmath75 , symmetry energy per nucleon at the saturation density of symmetric matter @xmath76 , slope @xmath77 and curvature @xmath78 of symmetry energy at symmetric matter saturation density ) , no functional model gives completely reliable extrapolations of nuclear masses in the neutron - rich region where experimental masses are not available .
this leads to a model dependence of the results , which is not related to any unknown nuclear matter parameter , but rather to the poor performance of the used hartree - fock or thomas - fermi approximations in the evaluation of nuclear masses , especially in the very neutron rich region .
very few nuclear functional models exist , with parameters fitted with the same degree of accuracy on infinite nuclear matter and to properties of finite nuclei @xcite , and even in this case the behavior of nuclear mass towards the dripline is subject to great uncertainties . anyway , since in the collapse applications the energetics of the gas plays a negligible role , the argument of consistency is not very important .
for this reason we have chosen to employ , whenever experimental measurements are missing , the duflo - zuker mass model @xcite which gives , within a microscopically inspired formalism , the present best reproduction of measured nuclear masses . for the unbound nucleon component the sly4 @xcite skyrme effective interaction is used . since we are here mainly interested in the region around the @xmath0 and @xmath1 , and in order to get an idea of the performance of the different mass models , fig .
[ fig : mass ] shows the evolution of the two - neutron separation energy for different elements around the magic neutron numbers @xmath0 and @xmath1 as predicted by the 10 parameter model of duflo and zuker@xcite and two other extremely successful and popular nuclear mass models , the bruxelles functional bsk22 @xcite and the finite range droplet model ( frdm ) by moller and nix @xcite , in comparison to experimental data from ref .
the overall dispersion of the models on the binding energy over the global mass table is 561 kev for dz10 , 580 kev for bsk22 , and 656 kev for frdm @xcite .
we can observe that all models appear approximately equivalent in the level of reproduction of experimental data .
their extrapolation in the neutron rich region however increasingly deviate from each other approaching the dripline .
in particular , a clear and strong slope change is visible at @xmath0 and @xmath1 for all elements including the lightest ones for the phenomenological models dz10 and frdm .
this indicates the presence of a very strong magic number , unmodified by the increasing neutron richness .
conversely the more microscopic bsk22 shows a more irregular behavior , and a certain quenching of magicity going towards the neutron dripline @xcite , especially for @xmath0 .
this shell quenching appears to be closely related to the treatment of the pairing interaction .
a proper inclusion of pairing correlations is certainly a very delicate issue , but one can at least expect that it should be better treated in a self - consistent hartree - fock - bogoliubov calculation such as bsk22 than in a phenomenological mass model .
these differences show the difficulty of extrapolation of nuclear masses far from stability .
) reached by the central element in the core - collapse of two progenitor stars with zero age main sequence masses @xmath79 and @xmath80 as reported in ref.@xcite .
simulations are from refs . @xcite . ]
exact values of temperature , baryon number density and proton fraction reached during core collapse depend on many ingredients and among others the chosen eos and weak interaction rates .
thus , without performing a simulation , the results can not be completely consistent .
we think , however , that it is sufficient to take typical values for the purpose of the present paper , where we aim to illustrate the possible impact of a modification of nuclear masses far from stability on the electron capture rates .
a detailed simulation is left for future work . to obtain such typical values , we consider here two different core collapse trajectories from refs .
@xcite , as reported by juodagalvis et al . in ref .
they correspond to the pre - bounce evolution for the central element of the star at an enclosed mass of 0.05 solar masses , in the case of a @xmath79 and a @xmath80 progenitor .
these simulations use rates from lmsh tables @xcite and the fully general relativistic , spherically symmetric agile- boltztran code with detailed neutrino radiation hydrodynamics . for further details on the simulations , see refs .
a more complete study would certainly require a complete and consistent simulation .
we expect , however , that the qualitative findings do not depend strongly on the quantitative details .
in addition , it was observed @xcite that the electron fraction profiles are well correlated with the density during the collapse phase , meaning that matter , independently of the exact position inside the star , will follow similar trajectories and that our example conditions taken from the central element are valid more generally . the thermodynamic conditions during the evolution in terms of temperature , baryon number density and proton fraction are shown in fig . [ fig : thermocond ] .
the vertical bar indicates the region in @xmath81 , where the experimental information on nuclear masses starts to be incomplete according the nuclear abundances obtained within our nse model employing experimental masses @xcite complemented with the dz10 mass table @xcite . for both progenitors
, this happens at densities higher than @xmath82 @xmath37 , and temperatures above @xmath83 mev .
we expect that the mass model has a considerable influence at this stage of the collapse . ) and heavy ( @xmath84 ) clusters obtained for the central element of the core - collapse with a @xmath79 ( thin blue lines ) and a @xmath80 ( thick magenta lines ) progenitor , see ref .
the temporal evolution is labeled via the baryon number density as in fig .
[ fig : thermocond ] .
the color coding is the same as in fig .
[ fig : thermocond ] . ] at any time during collapse , and for both collapse trajectories , the cluster mass distribution within the nse model , using experimental masses @xcite complemented with the dz10 mass table @xcite for the nuclei for which experimental mass measurements are not available , has a multi - peaked structure .
a first light cluster component exponentially decays up to a minimum around @xmath85 ; then the abundances increase and show one or several peaks in regions of the mass table which evolve according to the collapse time . to have a global glance of these distributions , fig .
[ fig : massfrac ] displays the mass fraction corresponding to nucleons , light clusters ( @xmath86 ) and heavier clusters ( @xmath84 ) .
the thermodynamic conditions are labeled by the baryon number density .
other mass models lead to very similar results , and we will keep dz10 as our fiducial mass model in the subsequent calculations .
we can see that the higher progenitor mass leads to a higher mass fraction for the lighter elements ( nucleons and light clusters ) , while heavy nuclei are largely dominant at all times with the less massive progenitor at this stage of the collapse .
the reason is the systematically higher temperature in the central element of the more massive star ( see fig .
[ fig : thermocond ] ) .
one can observe , too , that the behavior of the clusters abundances is non - monotonic for the conditions obtained with the higher mass progenitor .
this is due to the interplay between the increasing temperature during collapse , which favors light clusters with respect to heavy ones , and the increasing density leading to an opposite effect . )
nuclei produced in the central element of the core - collapsing @xmath87 and @xmath80 progenitors @xcite as a function of baryonic density .
the vertical bars correspond to the standard deviation of the distribution . for better readability ,
@xmath43 and @xmath44 data have been slightly displaced in density .
, title="fig : " ] ) nuclei produced in the central element of the core - collapsing @xmath87 and @xmath80 progenitors @xcite as a function of baryonic density .
the vertical bars correspond to the standard deviation of the distribution . for better readability ,
@xmath43 and @xmath44 data have been slightly displaced in density .
, title="fig : " ] the composition of heavy ( @xmath84 ) nuclei is further explored in fig .
[ fig : azn ] , which shows for the two trajectories the average and most probable neutron and proton numbers of these nuclei .
we can see that the two trajectories produce qualitatively similar patterns .
more precisely , by increasing the density larger clusters are produced .
increased densities imply , along the considered trajectories , increased neutron enrichment of stellar matter which , in turn , leads to increased neutron enrichment of nuclear clusters .
the standard deviation of the neutron and proton numbers distribution ( signaled by the vertical bars ) are never negligible meaning that treating matter composition within the single nucleus approximation ( sna ) would have produced erroneous results , as already acknowledged by the pioneering works @xcite . in both cases and over most of the explored density range average @xmath43 and @xmath44 numbers differ from the most probable ones .
one reason is that the distributions are not only broad but multi - peaked , too . indeed , very often the most probable nuclei are magic in @xmath43 .
magicity in @xmath44 is less frequent because of the smaller total number of protons .
the reduced number of protons with respect to neutrons explains also the systematically narrower distributions in @xmath44 . finally , the increasing width of the distributions in neutron number with increasing density arise due to the competition of a second magic number , in the present case @xmath8 .
the stability of the fragment production pattern is in agreement with the findings by ref .
@xcite , who have obtained a very similar multiplicity distribution pattern for a bunch of different trajectories with different progenitor models and eos .
we therefore conclude that the present simulation is reasonably representative of the generic evolution of the central element .
@xmath37 , @xmath88 mev , @xmath89 .
the solid lines mark the boundaries of experimental mass measurements .
the dotted lines mark magic numbers . ] to better understand the origin of this behavior , fig.[fig : ynz_t=2 ] displays the isotopic abundances corresponding to @xmath90 @xmath37 , @xmath88 mev , @xmath89 , _ i.e. _ to the time where the dispersion of the distribution of fig .
[ fig : azn ] starts to become non - negligible .
the distribution is centered around the @xmath0 neutron magic number , and the important width is due to the emergence of a second peak around @xmath1 .
this finding is in agreement with the simulations in ref .
@xcite and a similar effect was observed for the neutron star crust in refs .
@xcite . in particular , in ref .
@xcite it was shown that the overall variation of the electron fraction during the collapse is most sensitive to the electron capture rate on nuclei in the mass range 74@xmath9184 , particularly on @xmath2ni , @xmath92cu , and @xmath92zn close to the @xmath93 shell closure .
our findings confirm these results .
the two lines in fig.[fig : ynz_t=2 ] show the borders of the region where mass measurements exist , though with a variable degree of precision .
we can see that the most abundant nuclei lay just outside this border .
this means that their abundance , and therefore their preeminent role in the electron capture mechanism , relies on the extrapolation of the @xmath0 and @xmath1 shell closure far from stability , in a neutron rich region where mass measurements do not exist and spectroscopic informations are scarce and incomplete .
this means that small modifications in the nuclear binding energies of nuclei with masses not yet measured , but measurable in a near future , can change star matter composition and , consequently , all astrophysical quantities depending on it , notably weak rates .
this point will be explored in more detail in the following subsection . )
nuclei , heavy ( @xmath84 ) nuclei and the 30 ( dashed blue lines ) and 60 ( dashed cyan lines ) , respectively , most probable heavy nuclei .
sna - like ec rates ( see text ) are also plotted ( sna - mph ) .
the considered thermodynamical trajectories correspond to the central mass element of the @xmath94 and @xmath40-progenitors@xcite .
, title="fig : " ] ) nuclei , heavy ( @xmath84 ) nuclei and the 30 ( dashed blue lines ) and 60 ( dashed cyan lines ) , respectively , most probable heavy nuclei .
sna - like ec rates ( see text ) are also plotted ( sna - mph ) .
the considered thermodynamical trajectories correspond to the central mass element of the @xmath94 and @xmath40-progenitors@xcite .
, title="fig : " ] for this fiducial model , we now show the nse averaged electron capture ( ec ) rates on the different species @xmath95 , @xmath96 .
the two above mentioned core - collapse trajectories will be considered and l03 formulae ( see sec . [
sec : individual ] ) will be used .
let us first concentrate on the capture rates on heavy nuclei .
as it is well known @xcite , a very huge number of different nuclear species contribute to the total rate in all thermodynamics conditions .
this can be appreciated by limiting the rate calculation to the @xmath43 most probable nuclei , @xmath97 , where @xmath98 is the abundance of the @xmath99 most probable cluster .
comparing the result obtained with @xmath100 and @xmath101 , with the one corresponding to the whole distribution , one can see how important it is to properly account for the complete distribution of nuclear species . in the case of both progenitors
the 60 most probable nuclei never exhaust the average ec rate on heavy nuclei and , for instance , at @xmath102 @xmath37 they account for only 60% of @xmath103 . the inadequacy of the single nucleus approximation was recently stressed in ref.@xcite . in that work
it was shown that sizable differences in the collapse dynamics are obtained if the nse model is replaced with a more conventional model @xcite considering a single representative wigner - seitz cell for each thermodynamic condition , even if the same tm1 energy functional was employed in both models . however , in that work the individual rates were replaced by a single rate on the most probable cluster , using the simplified bruenn parametrization @xcite .
our results show that the complete nuclear distribution should also be used in the calculation of the rates .
if we replace the folding of the individual rates with the ec rate of the most probable nucleus weighted by the baryon number fraction bound in clusters , @xmath104 , the result ( lines with points ) is seen to very badly reproduce the complete folding result .
as one may see in fig .
[ fig : ec ] , at low densities @xmath105 generally underestimates the nse averaged ec rate and , because of structure effects related to the low temperatures at these densities , manifests a huge scattering . on the contrary , at higher densities and temperatures ( @xmath106 @xmath37 ) , @xmath105 largely overestimates the nse - averaged ec rate . obviously , different values are expected if instead of the l03 approximation , other prescriptions are employed for the individual ec rates , see the discussion in ref .
@xcite , too .
however , the general trend induced by the different averaging procedures should remain the same . concerning the ec rates on light nuclei , they are negligible at the beginning of the collapse but increase strongly , mainly due to the increasing temperature , and become dominant in the latest stage .
this underlines the importance of including microscopic calculation of ec rates for light nuclei , too @xcite .
the total rates are given by the sum of the different contributions , including electron capture on free protons ( dotted lines ) which however plays a minor role at all times .
comparing figs .
[ fig : thermocond ] , [ fig : azn ] and [ fig : ec ] , we note that for thermodynamic conditions where very neutron rich nuclei start to dominate , with masses which are not experimentally known , they still represent the major electron capture source .
this qualitative picture is independent of the progenitor mass . from a more quantitative point of view , the relative importance of heavy clusters is higher for the lower mass progenitor , essentially because of the lower temperatures reached during the collapse .
it is well known in the recent nuclear structure literature that even major shell closures can be quenched far from stability , see e.g. @xcite . a very clear evidence exists for the @xmath107 magic number @xcite , which corresponds to a huge gap for decreasing proton number up to @xmath108si , and suddenly disappears in the next even - even isotope @xmath109 mg .
this is partly due to the modification of single - particle energies far from stability , with the consequence that new magic numbers can appear corresponding to strongly deformed configurations .
however the main reason of the modification of magicity relies on effects which go beyond the naive single particle shell model .
first , the effect of the proton - neutron residual interaction in nuclei with strong asymmetry is decreased ; moreover , correlations play an increasing role which makes the very concept of shell closure less relevant . as a consequence ,
the main effect is a quenching of the shell gap , even if secondary new gaps can appear in very localized regions of the nuclear chart .
the possible modification of the @xmath0 and @xmath1 shell gaps far from stability is the object of intense theoretical and experimental research in nuclear structure ( e.g. @xcite ) . here
, we do not have the ambition to model this phenomenon , but simply analyze the modifications in the matter composition and associated electron capture rates , which would be induced by the expected shell quenching .
a similar idea was proposed in ref .
@xcite , who showed that a modified mass formula built to incorporate the possibility of shell quenching , has a striking impact on canonical calculations of the r - process .
we follow a similar strategy as in ref .
@xcite and introduce a modified expression for the binding energy in the following form : @xmath110 where @xmath111 and @xmath112 stand for the experimental binding energy @xcite and , respectively , predictions of dz10 mass model @xcite .
@xmath113 ( resp . , @xmath114 ) and @xmath115 ( resp .
@xmath116 ) correspond to the most neutron - rich and , respectively , most neutron - poor nucleus with @xmath42 nucleons for which experimental masses ( resp . , predictions of dz10 ) exist .
@xmath117 is a simple liquid - drop binding energy calculated according to @xmath118 with @xmath119 , @xmath120=15.62 mev , @xmath121=17.8 mev , @xmath122=29 mev , @xmath123=38.5 mev , @xmath124=0.7 mev and @xmath125 mev for even - even ( + ) and , respectively , odd - odd nuclei ( - ) . finally we introduce a smearing function depending on the parameters @xmath126 and @xmath127 which determine how sudden the shell quenching is supposed to be : @xmath128.\end{aligned}\ ] ] small values of @xmath126 correspond to maximum quenching , while in the limit @xmath129 we recover the dz10 functional form , which predicts preserved magic numbers up to the dripline . the behavior of the modified functional on the binding energy is illustrated in fig . [ fig : bmodif ] for different arbitrary values of the smearing parameter @xmath130 and @xmath131 .
a selected number of nuclei is displayed as a function of neutron number .
these elements have been chosen since they are strongly populated in the later phase of the collapse . as observed before , the dz10 model shows an extremely pronounced shell closure at @xmath0 and @xmath1 for all elements including the exotic ones like @xmath132ti and @xmath133kr . in the modified expression , eq .
( [ eq : bmod ] ) , the gap is quenched far from stability , and the quenching is more or less pronounced depending on the choice of the parameter @xmath126 . the effect of the modification of the mass formula on the distribution of nuclei is shown in fig.[fig : distrimodif ] for three different representative conditions during the collapse , taken from the 25@xmath134 progenitor @xcite . obviously , the effect of shell quenching is to reduce the size of the magic peaks and favor open shell nuclei , thus leading to a wider isotopic distribution .
corresponding to two different scenarios of shell quenching ( see text for details ) .
, title="fig : " ] corresponding to two different scenarios of shell quenching ( see text for details ) .
, title="fig : " ] impact of nuclear binding energy on nuclear abundances : the distribution of clusters with a given neutron number is shown .
the considered thermodynamic conditions ( @xmath3 [ mev ] , @xmath4 [ @xmath37 ] , @xmath5 ) are mentioned on each panel , corresponding to three different times in the evolution of the central element of the collapse with a 25 @xmath134 progenitor @xcite .
the same prescriptions for the binding energies as in fig .
[ fig : bmodif ] are used .
[ fig : distrimodif ] ( dashed lines ) and a @xmath40 ( full lines ) progenitor @xcite .
the temporal evolution is labeled by the baryon number density as before .
the averaged rate is calculated only on heavy nuclei ( @xmath84 ) in the upper panel and on all nuclei in the lower panel .
different quenching factors @xmath126 are considered ( see text ) . , title="fig : " ] ( dashed lines ) and a @xmath40 ( full lines ) progenitor @xcite
. the temporal evolution is labeled by the baryon number density as before .
the averaged rate is calculated only on heavy nuclei ( @xmath84 ) in the upper panel and on all nuclei in the lower panel .
different quenching factors @xmath126 are considered ( see text ) .
, title="fig : " ] our final result about the impact of shell quenching towards the drip - line on nse - averaged electron capture rate is shown in fig .
[ fig : ec_modif ] .
several magicity quenching recipes have been considered @xmath136 and @xmath137 for each of the two progenitors .
their predictions are plotted in terms of relative deviations with respect to the fiducial model .
panel ( a ) corresponds to nse - averaged ec rates on heavy nuclei . independently on the stiffness of magicity quenching and progenitor mass , the same pattern is obtained . over a certain time
after neutron - rich nuclei start to be copiously populated shell quenching induces an increase of the ec rate .
this is due to a significant abundance increase for non - magic nuclei and a moderate abundance decrease of the n=50 magic nuclei . in this regime shell quenching on the ec rates on heavy nuclei
is more pronounced , and lasts for a longer time , in the case of the 25@xmath134 progenitor . toward the end of the considered trajectories ,
the opposite effect is obtained . here
shell quenching is responsible for a reduction of one order of magnitude in the abundance of n=50 and n=82 magic nuclei , not compensated by the abundance increase of non - magic nuclei and the corresponding increase in the @xmath13-value . in the two regimes ,
the amplitude of the effect depends on both quenching parameter @xmath126 and progenitor mass and may amount up to 30% .
panel ( b ) shows that , despite the fact that heavy nuclei represent only a fraction of the whole mass and more isospin - symmetric light nuclei are not affected by shell quenching , modification of ec rate on heavy nuclei survives in the inclusive rate . up to @xmath138 @xmath37 , where @xmath139 ( see fig .
[ fig : massfrac ] ) , @xmath140 . for the highest considered densities
the overall modification of ec rates is opposite to that seen for heavy clusters .
more precisely , shell quenching here points toward an increase of ec .
this happens because of a more favorable production of light clusters .
in this paper we have examined the consequence of a possible quenching of the @xmath0 and @xmath1 shell closures on the ec rates during core collapse . as basis of the analysis
, we have considered the same typical thermodynamic conditions as in ref .
@xcite . they correspond to the pre - bounce evolution of the central element of the star obtained within a core - collapse simulation using two different progenitors , a 15@xmath134 and a 25@xmath134 one , from refs .
@xcite . in agreement with ref .
@xcite we find that the properties of very exotic nuclei around these two shell closures is a key microscopic information to predict the evolution of the electron fraction during collapse .
we have pointed out that a quenching of these shell closures compared with the popular dz10 mass model @xcite considerably affects the nuclear distribution , and consequently the ec rates during collapse . using the standard l03 formalism for rates on individual nuclei ,
we have analyzed the modification of nse average ec rates for different scenarios of shell quenching .
depending on the progenitor mass and stiffness of shell quenching modifications of ec rates of up to 30% have been obtained .
we expect that such effects , once consistently included in the time dependent evolution of the collapse , have a sizable effect on neutrino emissivity and on the enclosed mass at bounce .
we should mention that the l03 approximation might be questionable in this exotic region far away from the stable @xmath11-shell nuclei where these formulae were adjusted .
we expect , however , that more precise rates do not change our qualitative findings .
the reason is that the ec rates are considerably suppressed for nuclei with a pronounced shell closure due to the energy gap in the single particle states which makes a transition more difficult . in this work
the quenching effect is governed by the parameter @xmath141 whose value is chosen in an arbitrary way .
more sophisticated calculations in microscopic theories , such as modern interaction configuration shell model calculations in the relevant nuclear chart region @xcite are currently underway , and will give essential information on the effective importance of the shell quenching .
however , the last word will be clearly given by phenomenology .
new precise mass measurements at the edges of the known isotopic table could allow for a much better extrapolation towards the neutron rich region in a very near future . finally , it is important to stress that the modification of nuclear structure far from stability is expected to influence not only the nuclear mass , but also the individual ec probabilities , which should be consistently calculated together with the mass model and the energy functional describing the self - interaction of unbound particles .
dedicated qrpa calculations are in progress to achieve this aim @xcite .
this work has been partially funded by the sn2ns project anr-10-blan-0503 and by newcompstar , cost action mp1304 .
r. r acknowledges partial support from the romanian national authority for scientific research under grants pn - ii - id - pce-2011 - 3 - 0092 and pn 09 37 01 05 and kind hospitality from lpc - caen and luth - meudon .
h. schatz , a. aprahamian , j. gorres , et al .
phys . rept .
* 294 * , 167 ( 1998 ) .
s. gupta , e. f. brown , h. schatz , p. moeller and k. l. kratz , astrophys .
j. * 662 * , 1188 ( 2007 ) .
h. schatz _ et al .
_ , nature * 505 * , 7481 , 62 ( 2014 ) .
g.audi , m. wang , a. h. wapstra , f. g. kondev , m. maccormick , x. xu , and b. pfeiffer , chinese physics *
c36 * , 1287 ( 2012 ) ; m. wang , g. audi , a. h. wapstra , f. g. kondev , m. maccormick , x. xu , and b. pfeiffer , chinese physics *
c36 * , 1603 ( 2012 ) ; http://amdc.impcas.ac.cn/evaluation/data2012/data/nubase.mas12 . | the importance of microphysical inputs from laboratory nuclear experiments and theoretical nuclear structure calculations in the understanding of the core collapse dynamics , and the subsequent supernova explosion , is largely recognized in the recent literature . in this work ,
we analyze the impact of the masses of very neutron rich nuclei on the matter composition during collapse , and the corresponding electron capture rate . to this aim
, we introduce an empirical modification of the popular duflo - zuker mass model to account for possible shell quenching far from stability , and study the effect of the quenching on the average electron capture rate .
we show that the preeminence of the @xmath0 and @xmath1 closed shells in the collapse dynamics is considerably decreased if the shell gaps are reduced in the region of @xmath2ni and beyond . as a consequence ,
local modifications of the overall electron capture rate up to 30% can be expected , with integrated values strongly dependent on the stiffness of magicity quenching and progenitor mass and potential important consequences on the entropy generation , the neutrino emissivity , and the mass of the core at bounce .
our work underlines the importance of new experimental measurements in this region of the nuclear chart , the most crucial information being the nuclear mass and the gamow - teller strength .
reliable microscopic calculations of the associated elementary rate , in a wide range of temperatures and electron densities , optimized on these new empirical information , will be additionally needed to get quantitative predictions of the collapse dynamics . |
it is common knowledge that the scalar potential @xmath3 of any charge distribution @xmath4 can be calculated ( using the gaussian absolute system of units ) by the relation ( see , e.g. , @xcite , chapter 2 ) @xmath5 in the spherical coordinate system @xmath6 the @xmath3 potential can be conveniently found in many cases by expansion in terms of spherical harmonics @xmath7 ( see , e.g. , @xcite ) .
the potential at a point is a sum of actions of all the charges involved .
therefore , if we are interested in the potential at a point inside a charge distribution , we will have to split the region of integration in eq .
( [ 1 ] ) into two parts by a sphere of radius @xmath8 centered at the pole @xmath9 . as a result
, we come to equalities of two types , for @xmath10 and for @xmath11 . here
@xmath8 refers to the point with the potential we are interested in , and @xmath12 is the point over which integration will be performed .
accordingly , the part of the potential formed by charges with coordinates @xmath13 we shall denote with @xmath14 , and the other part , produced by charges with coordinates @xmath15 , with @xmath16 .
the expressions for @xmath14 and @xmath16 written in the form appropriate for us here can be found , for instance , in @xcite and are reproduced below : @xmath17 where @xmath18 is a multipole moment of order @xmath19 : @xmath20 @xmath21 where @xmath22 : @xmath23 the functions @xmath18 and @xmath22 depend on @xmath8 on the upper and lower limits of integration as on a parameter .
the potential at a given point @xmath24 is a sum of the potentials calculated for @xmath10 and @xmath11 : @xmath25 . the spherical functions @xmath7 can be written as ( see , e.g. , @xcite ) : @xmath26 in this expression , the integers @xmath27 satisfy the conditions @xmath28 , @xmath29 ; @xmath30 for @xmath31 , @xmath32 for @xmath33 .
@xmath0 denotes here the associated legendre polynomial .
the vector potential @xmath34 of any current distribution @xmath35 in vacuum can be calculated from the relation ( see , e.g. , @xcite , chapter 2 ) @xmath36 where @xmath37 is the electrodynamic constant .
relation ( [ 7 ] ) resembles in its form equality ( [ 1 ] ) , the difference being that integration is performed here not over the scalar but rather over the vector function @xmath35 .
if one wanted to employ the above method of potential calculation with the use of spherical harmonics , one would have to apply all the calculations to each component separately , with subsequent vector addition of the results thus obtained .
this approach may meet with considerable difficulties .
on the other hand , if the currents have axial symmetry , one can suggest a fairly simple and straightforward method of calculation of the vector potential which reduces essentially to calculating an integral of one scalar function only .
assume a system of currents with a volume density @xmath35 which have axial symmetry .
denote the axis of symmetry of the system with @xmath38 . introduce a spherical coordinate system @xmath6 such that the @xmath39 axis coincides with the symmetry axis @xmath38 .
denote the unit vectors along the @xmath6 axes by @xmath40 .
the @xmath41 plane will be called the equatorial plane .
the system being symmetric , the currents do not depend on the @xmath42 coordinate and have only one component along it : @xmath43 .
this means that all @xmath35 are circular currents , and that the vector of any element of the current lies in the plane parallel to the equatorial plane ( we shall call it the @xmath44 plane ) .
the symmetry of the current system suggests that the vector potential @xmath34 is also axially symmetric , i.e. , the magnitude of the vector potential @xmath45 does not depend on coordinate @xmath42 .
but this means that the magnitude of the vector potential @xmath45 can be determined on any plane passing through the symmetry axis @xmath38 .
the simplest way appears to determine the potential on the plane passing through the @xmath38 axis and the coordinate @xmath46 , and it is this what we are going to do .
call this plane the @xmath47 plane .
calculate the vector potential at a point @xmath24 lying on the @xmath47 plane .
the coordinates of this point are @xmath48 .
it is convenient for the purpose of this analysis to pass through this point a plane @xmath49 parallel to the equatorial plane .
consider our situation in more detail .
vector potential at a point @xmath24 ( lying on plane @xmath47 ) is produced by currents flowing throughout the space of interest here .
consider the potentials created by the current elements @xmath50 .
figure 1 illustrates our case . in fig .
1 , the origin of the spherical system of coordinates is at point @xmath9 , the @xmath38 axis points directly at us , and the dashed line visualizes one of the circular current trajectories .
this trajectory lies in the @xmath44 plane .
the @xmath49 and @xmath44 planes are parallel but do not coincide .
the direction of the @xmath42 axis at point @xmath48 is specified in the figure .
we are looking for the potential at this point .
the @xmath42 axis at point @xmath48 is perpendicular to the @xmath47 plane .
the @xmath8 axis lies in the @xmath47 plane . the dashed line @xmath51 on the @xmath44 plane is perpendicular to plane @xmath47 , i.e. , parallel to the @xmath42 axis at point @xmath48 .
figure 1 shows two elements of the circular current with densities @xmath52 and @xmath53 .
the current element @xmath52 produces at point @xmath48 on plane @xmath47 the potential @xmath54 . as seen from vector equality ( [ 7 ] ) , this potential is directed in space in the same way as the current element @xmath52 .
this means that the angles between a direction in space ( for instance , straight lines perpendicular to plane @xmath47 ) and vectors @xmath52 and @xmath55 should be equal ( in fig .
1 , these are the @xmath56 angles ) .
the vector of the current element @xmath52 is confined to plane @xmath44 , which is parallel to the equatorial plane . as seen from expression ( [ 7 ] ) , this means that the element of the vector potential @xmath55 generated by the current element @xmath52 should likewise lie fully in the plane parallel to the equatorial plane , i.e. , in plane @xmath49 ( the @xmath49 and @xmath44 planes do not coincide ) .
therefore , the @xmath55 potential can be resolved into two components , @xmath57 and @xmath58 in the @xmath49 plane .
these components are shown in fig . 1 : @xmath59 and @xmath60 , where the angle @xmath56 is the coordinate of the current element @xmath52 .
since , however , the current is axially symmetric , there will always be an element of current @xmath61 equal in magnitude to the element @xmath62 and located at the same distance from point @xmath48 as the @xmath62 element .
the coordinates of this current element are @xmath63 .
it produces at the same point @xmath48 the potential @xmath64 . resolving potential @xmath64 in the @xmath49 plane into two components , @xmath65 and @xmath66 , we see immediately that the components @xmath57 and @xmath65 cancel , while @xmath58 and @xmath66 add ( see fig .
1 ) . thus , after integration over all elements of current @xmath50 the potential @xmath34 will have only one component , the one along the @xmath42 axis , left . but
this implies that in determination of the @xmath34 potential one may , rather than taking into account the whole potential @xmath67 produced by the current element @xmath50 , restrict oneself to the projection @xmath68 of this potential onto the @xmath42 axis : @xmath69 , where @xmath56 is the coordinate of the element of current @xmath50 .
said otherwise , the potential @xmath68 is generated not by the current element @xmath50 as a whole but by the projection of this element on the line @xmath51 parallel to vector @xmath42 at point @xmath48 .
we can now , on replacing vector @xmath35 in eq .
( [ 7 ] ) with a scalar @xmath70 , calculate the magnitude of the vector potential @xmath71 the potential at any point @xmath24 can now be written as @xmath72 because expression ( [ 8 ] ) contains scalar quantities , in calculation of the integral one may resort to expanding the integrand in spherical harmonics .
note , however , that because expression ( [ 8 ] ) contains @xmath73 as a factor , one may conveniently use in expressions for @xmath7 in place of exponential functions employed customarily , trigonometric ones ( see .
e.g. , @xcite , chapter 21 ) . in this case
one will have to replace expression ( [ 6 ] ) with the following relations : @xmath74 @xmath75 @xmath76 this involves a minor change in the form of expansion of the function . for @xmath10 :
@xmath77 where the multipole moments @xmath78 and @xmath79 have the form @xmath80 @xmath81 for @xmath11 : @xmath82 with the multipole moments @xmath83 and @xmath84 @xmath85 @xmath86 the function @xmath73 entering the expression for @xmath87 also belongs to the system of functions @xmath88 ( see expression ( [ 10 ] ) ) . therefore , because of the functions @xmath89 and @xmath90 being orthogonal , part of the multipole moments will vanish in the course of integration of expressions ( [ 13 ] ) , ( [ 14 ] ) and ( [ 16 ] ) , ( [ 17 ] ) ( orthogonality in index @xmath1 ) .
the only terms left will be those with @xmath2 , i.e. , the terms @xmath91 and @xmath92 . as a result , the double series ( [ 12 ] ) and ( [ 15 ] ) will become single series and look as @xmath93 @xmath94 the next step appears to be substitution in these series of the expression for @xmath95 from formula ( [ 10 ] ) .
we have to remember , however , that we are looking for the potential on the @xmath46 plane only .
for all other values of the @xmath42 coordinate , the relations we have derived will yield wrong values of the potential .
therefore , prior to substituting the expression for @xmath95 from relation ( [ 10 ] ) , we will have to zero the coordinate @xmath42 .
we obtain naturally @xmath96 .
now the potentials take on the form @xmath97 where the multipole moments @xmath98 are @xmath99 @xmath100 with the multipole moments @xmath101 @xmath102 in these formulas , @xmath103 is the circular current density , and @xmath104 and @xmath105 are the parts of the multipole moments left after integration over the @xmath42 coordinate ( truncated multipole moments ) .
the vector potential at any point in space can be written as @xmath106 \mathbf{n}_\varphi,\ ] ] whence we finally come to this is the final expression for the vector potential @xmath34 . as evident from eq .
( [ 25 ] ) , vector potential @xmath34 has only one component , @xmath107 .
now we impose an additional constraint , namely , that the currents are symmetrical with respect to the equatorial plane ( which quite often is the case ) . in this case ,
expression ( [ 25 ] ) may contain only functions @xmath108 symmetric about the equator , i.e. , @xmath109 functions .
now expression ( [ 25 ] ) for the vector potential has to be replaced by @xmath110 it may be of interest to consider a particular relevant case .
consider axially symmetric currents with a @xmath111 dependence on the @xmath112 angle .
this case becomes realized , for instance , if the currents are generated by rotation of charges , with all the charges in a layer of thickness @xmath113 distributed uniformly over the angle @xmath112 and rotating with the same angular velocity .
the @xmath111 function belongs to the system of associated legendre functions .
therefore , the @xmath108 functions being orthogonal with respect to index @xmath114 , the larger part of the terms in the @xmath104 and @xmath105 expansions ( see expressions ( [ 21 ] ) and ( [ 23 ] ) ) will vanish in calculation , with only @xmath115 and @xmath116 terms with index @xmath117 retained . in this case , the potential will come out as a sum of two terms only : @xmath118
magnetic field @xmath119 can be calculated with the general expression @xmath120 .
the vector potential of axially symmetric currents has only one component , @xmath107 : @xmath121 . with this in mind , and using the general expression for @xmath122 in the spherical coordinate system (
see , for instance , @xcite , chapter 6 ) @xmath123\mathbf{n}_r+\ ] ] @xmath124 \mathbf{n}_\vartheta+ \left[\frac{1}{r}\frac{\partial(ra_\vartheta)}{\partial r}-\frac{1}{r } \frac{\partial a_r}{\partial\vartheta}\right]\mathbf{n}_\varphi,\ ] ] we finally come to if the currents are symmetric about the equator , expression ( [ 30 ] ) , similar to formula ( [ 26 ] ) , will contain only equatorially symmetric functions @xmath109 , and summation over @xmath132 should be performed from @xmath133 to @xmath134 . to illustrate the use of this expression ,
consider the following simple problem : rotation of a charged sphere about one of the diameters .
let the sphere of radius @xmath136 , with charge @xmath137 distributed uniformly over its surface , rotate about the @xmath38 axis with an angular velocity @xmath138 .
calculate the magnetic field and the vector potential inside and outside the sphere .
@xmath144 where @xmath145 is the magnetic moment of the system .
one can readily verify that these expressions are identical . in monographs
@xcite and @xcite it is proposed to solve this problem by differential calculus .
our approach , by contrast , has yielded the answers to it by straightforward calculation . | a solution is proposed for finding the vector potential and magnetic field of any distribution of currents with axial symmetry . in this approach , the magnetic field and the vector potential
are looked for not by solving a differential equation but rather through straightforward calculation of integrals of one scalar function .
the solution is expressed in terms of the associated legendre polynomials @xmath0 with the index @xmath1 of the legendre polynomials assuming one value only , @xmath2 .
the solution has the form of a series , with the coefficients of the polynomials being combinations of multipole moments .
* key words : * electrodynamics , vector potential , spherical harmonics , legendre polynomials , magnetic field . |
both the author @xcite and paolo salvatore @xcite have constructed actions of the operad of little @xmath0-cubes on the ` framed ' long knot spaces @xmath1 defined in @xcite which extend the homotopy - associative product given by the connect - sum operation .
computations of the homotopy - type @xcite and homology @xcite of @xmath2 suggest in several ways ( some more vague than others ) that the spaces @xmath1 may admit an action of the operad of framed little @xmath0-discs . a concrete example : provided
such an action exists , it would allow for a more compact and ` operadic ' description of the homotopy - type of @xmath2 than the one given in @xcite .
the author has made several attempts to construct an action of the operad of framed @xmath0-discs @xmath3 on @xmath1 in the spirit of ` little cubes and long knots ' @xcite , as of yet without success .
salvatore constructs the spaces @xmath1 for @xmath4 as the totalisation of a multiplicative operad , further developing the work of sinha @xcite .
the operad salvatore uses is equivalent to the operad of framed little @xmath5-discs @xmath6 .
theorems of mcclure and smith imply that the totalisation of a multiplicative operad is a module over the operad of @xmath0-cubes @xcite .
motivated by these constructions , salvatore asked if the operad of framed discs is cyclic .
the geometric techniques used in the author s failed attempts to construct an action of @xmath3 directly on @xmath1 turn out to answer his question . in proposition
[ cbprop ] , @xmath7 is shown to be equivalent in the sense of fiedorowicz @xcite to the operad of ` conformal @xmath8-balls ' @xmath9 . in theorem
[ mainthm ] , @xmath9 is shown to be cyclic . this paper follows the conventions of the book ` operads in algebra , topology and physics ' by markl , shnider and stasheff @xcite .
let @xmath10 be the unit sphere in @xmath11 .
consider @xmath12 .
let @xmath13 and @xmath14 be the north and south poles of @xmath10 respectively .
let @xmath15 be the ` southern ' hemisphere in @xmath10 , meaning @xmath16 and @xmath17 .
given a point @xmath18 , identify the tangent space @xmath19 with @xmath20 via stereographic projection from @xmath21 .
since stereographic projection is a conformal diffeomorphism , one can consider a conformal affine - linear transformation of @xmath19 to be a conformal diffeomorphism of @xmath20 and further extend it to a conformal diffeomorphism of @xmath10 .
let @xmath22 be the group of orientation - preserving conformal diffeomorphisms of @xmath10 .
let @xmath23 be the subgroup corresponding to the conformal affine - linear transformations of @xmath19 .
let @xmath24 be the subgroup consisting only of translations in @xmath19 .
let @xmath25 denote the group of orientation - preserving isometries of @xmath10 .
let @xmath26 be defined as @xmath27 .
define @xmath28 to be the scalar multiples of the identity on @xmath19 . for @xmath29
, @xmath22 is known to be a @xmath30-dimensional lie group , isomorphic to the group of hyperbolic isometries of hyperbolic @xmath5-space .
the isomorphism is given by restriction to the projectivised light - cone of the minkowski model .
stated another way , for any @xmath18 , @xmath22 is generated by the subgroups @xmath25 , @xmath31 and @xmath32 .
moreover , the stabiliser of @xmath18 under the action of @xmath22 is @xmath33 . for details ,
see schoen and yau s book @xcite
vi theorem 1.1 .
let @xmath34 denote rotation by 180-degrees around @xmath35 .
a @xmath36-tuple @xmath37 will be called @xmath38 little conformal @xmath8-balls if @xmath39 for all @xmath40 , and if the interior of @xmath41 is disjoint from @xmath42 for all @xmath43 .
the set of all such @xmath36-tuples is denoted @xmath44 .
there are maps @xmath45 defined by sending @xmath46 to @xmath47 @xmath48 .
consider @xmath44 to be a subspace of @xmath49 .
then there is a natural right action of @xmath50 on @xmath44 .
the above maps together with the actions of @xmath50 for @xmath51 endow @xmath9 with the structure of an operad , hereby named the operad of conformal @xmath8-balls .
observe that for @xmath29 the operad of framed little @xmath8-discs @xmath7 is the suboperad of @xmath9 such that the north pole is fixed by each disc : @xmath52 for all @xmath53 .
* @xmath66 is translation by @xmath67 .
* @xmath68 is translation by @xmath69 .
* @xmath70 fixes both @xmath71 and @xmath72 , so it is an element of @xmath73 , and so there is a unique solution to the equation @xmath74
. given @xmath75 , let @xmath76 be as above .
let @xmath77 for @xmath78 $ ] be the straight line in @xmath79 from @xmath80 to the identity .
notice that @xmath81 for all @xmath40 and @xmath82 .
for all @xmath78 $ ] there exists @xmath83 $ ] such that if @xmath84 denotes scaling by @xmath85 in @xmath86 , then @xmath87 . for
each @xmath78 $ ] choose @xmath85 to be the maximal possible such element .
this makes @xmath85 a continuous function of @xmath82 .
observe @xmath88 .
one of the main roles of the operad of framed discs @xmath7 is that it detects when a space @xmath93 has the homotopy - type of an @xmath8-fold loop space @xmath94 where @xmath95 is an @xmath96-space @xcite . by the work of fiedorowicz @xcite
, the above proposition implies that @xmath9 also detects @xmath8-fold loop spaces over @xmath96-spaces .
we define a right @xmath97-action on @xmath44 which extends the @xmath50 action . here
@xmath97 is the permutation group of @xmath98 and @xmath50 is the stabiliser of @xmath99 .
given @xmath100 and @xmath101 , let @xmath102 .
juxtaposition is interpreted as composition of functions , where @xmath101 is considered to be a an element of @xmath49 .
more explicitly , @xmath103 where we interpret @xmath104 as @xmath105 , if it appears . this is a right action of @xmath97 , which can be verified by a quick calculation : @xmath106 the three axioms of a cyclic operad are verified below . axiom ( i ) @xmath107^{\tau_1 } & & { { \mathcal{cb}}}_n(1 ) \\ & 1 \ar[ul]^-{\eta } \ar[ur]_-{\eta } & } \ ] ] states that the transposition @xmath108 fixes the identity element of @xmath109 , which can be checked : @xmath110 . axiom ( ii )
states that the diagram below commutes : @xmath111 ^ -{\circ_1 } \ar[d]^{\tau_i \times \tau_j } & { { \mathcal{cb}}}_n(i+j-1 ) \ar[dd]^{\tau_{i+j-1 } } \\ { { \mathcal{cb}}}_n(i ) \times { { \mathcal{cb}}}_n(j ) \ar[d]^{s } & \\ { { \mathcal{cb}}}_n(j ) \times { { \mathcal{cb}}}_n(i ) \ar[r]^{\circ_i } & { { \mathcal{cb}}}_n(i+j-1 ) } \ ] ] which we check by going around the diagram two ways , first the clockwise direction .
@xmath112 axiom ( iii ) states that the diagram below commutes : @xmath113^{\circ_i } \ar[d]^{\tau_i \times 1 } & { { \mathcal{cb}}}_n(i+j-1 ) \ar[d]^{\tau_{i+j-1 } } \\ { { \mathcal{cb}}}_n(i ) \times { { \mathcal{cb}}}_n(j ) \ar[r]^{\circ_{i-1 } } & { { \mathcal{cb}}}_n(i+j-1 ) } \ ] ] which we check by going around the diagram both ways , first clockwise .
@xmath114 the operad of little @xmath8-cubes ( or unframed little @xmath8-discs ) is known not to be cyclic provided @xmath29 is even , since its homology is not cyclic .
a proof appears in proposition 3.18 of @xcite .
@xmath115 and @xmath3 were already known to be cyclic before the publication of this article @xcite .
the proof above can be adapted to give a coherent proof that @xmath9 is cyclic for all @xmath116 , provided one defines @xmath117 to be the group of orientation - preserving isometries of hyperbolic @xmath0-space , restricted to the circle at infinity .
one defines @xmath118 to be rotation by @xmath119 degrees about the centre of the circle , and the proof proceeds as above . | the operad of framed little discs is shown to be equivalent to a cyclic operad .
this answers a conjecture of salvatore in the affirmative , posed at the workshop ` knots and operads in roma , ' at universit di roma `` la sapienza '' in july of 2006 . |
the alkaline earth subnitrides x@xmath0n ( x = ca , sr , and ba ) in their bulk form can be synthesized by a direct solid - state reaction of the elements or through chemical reduction of the corresponding sesquinitrides@xcite .
their structures consist of alternating abc - stacked ( x - n - x ) hexagonal layers , which are in the @xmath6 space group ( anti - cdcl@xmath0-type ) with a high @xmath7 ratio , where @xmath8 and @xmath9 are the in - plane and out - of - plane unit cell dimensions@xcite .
each ( x - n - x ) layer is closely packed . for example , the [ ca@xmath0n ] layer has a thickness of 2.51 which is much smaller than that of the ordinary fcc ca ( 111 ) layers ( 3.18 ) .
this is typically attributed to the ionic bounding in the layer . meanwhile
, the separation between two neighboring layers is relatively large , leading to 2d confined spaces for the anionic electrons@xcite . in this work
, we investigate the properties of monolayer and few - layer thin films of these subnitrides ( here one monolayer ( ml ) refers to one ( x - n - x ) unit layer ) .
since large interlayer separations ( @xmath10 ) usually signal a possible role of long - range van der waals interaction , the van der waals corrections was included in the first - principles calculation@xcite .
the calculation details are described in the method section . as an example , the crystal structure of 3-ml x@xmath0n ( x = ca , sr ) is schematically shown in fig.[fig1]a .
the optimized lattice parameters for isolated ca@xmath0n and sr@xmath0n thin films with thickness from 1-ml to 5-ml are listed in table [ lattice ] .
one observes that despite slight variations with respect to the number of layers , the calculated in - plane lattice constants , layer thicknesses , and interlayer distances are close to the experimental values of the corresponding bulk structures ( for bulk ca@xmath0n , @xmath11 , layer thickness@xmath12 , and layer separation@xmath13 ; for bulk sr@xmath0n , @xmath14 , layer thickness@xmath15 , and layer separation@xmath16)@xcite .
generally , the interlayer spacing is much larger than the layer thickness by about 50% .
sr@xmath0n has a larger interlayer spacing ( around 4 ) than ca@xmath0n due to the larger atomic number of sr .
the dynamical stability of 1-ml ca@xmath0n has been studied in ref . .
here we find that few - layer ca@xmath0n and sr@xmath0n ( including sr@xmath0n monolayer ) are also dynamically stable through analysis of their phonon spectra .
in contrast , the similar structures of 1-ml and 2-ml ba@xmath0n are not stable due to the presence of imaginary frequencies in their phonon spectra .
therefore , we only focus on ca@xmath0n and sr@xmath0n in the present work . as representative examples ,
the phonon dispersions of monolayer ca@xmath0n and sr@xmath0n are shown in fig.[fig2 ] .
the absence of imaginary frequencies in the whole brillouin zone demonstrates that the dynamical stability of the corresponding structures .
the in - plane and out - of - plane transverse acoustic modes are not degenerate due to the structural anisotropy which is typical for layered materials .
more importantly , while the in - plane transverse acoustic modes have a linearly dependence on their wavevectors in the vicinity of @xmath17-point , the out - of - plane acoustic ( za ) phonons exhibit a quadratic dispersion around @xmath17-point .
the parabolic dispersion of za modes is a characteristic feature of layered materials , which is consistent with the macroscopic elastic theory of thin plates@xcite .
this feature has been frequently observed in other layered 2d materials , such as graphene@xcite , layered transition metal dichalcogenides@xcite , and phosphorene@xcite .
we then calculate the surface energies of 1-ml to 5-ml ca@xmath0n and sr@xmath0n and the results are plotted in fig.[fig3 ] . because of their layered structure and the large interlayer separations , the surface energies are relatively low .
the calculated values are comparable to the surface energy of graphene ( @xmath18mev/@xmath19)@xcite . previous calculations on 1-ml ca@xmath0n have suggested its possibility to be mechanically exfoliated from the bulk . from our result in fig.[fig3 ] , the surface energy decreases with the increasing number of layers .
hence few - layer ca@xmath0n would be more easily exfoliated compared with monolayer ca@xmath0n , as naturally expected .
sr@xmath0n has an even larger interlayer spacing .
surface energy for 1-ml sr@xmath0n is about @xmath20mev/@xmath19 , and it decreases to about @xmath21mev/@xmath19 for 5-ml . with the same number of layers , sr@xmath0n s surface energy is smaller than that of ca@xmath0n by about @xmath22mev/@xmath19 .
this implies that sr@xmath0n monolayer and few - layers could be more easily obtained by mechanical exfoliation using scotch tapes or afm tips .
bulk ca@xmath0n and sr@xmath0n have conduction electrons confined in the 2d interlayer regions playing the role of anions@xcite .
one key question is whether these anionic electron layers are maintained in their thin film structures .
we find that the answer is positive . in fig.[fig4 ] , we show the electronic band structures of x@xmath0n ( x = ca , sr ) from 1-ml to 3-ml .
first , one notices that all these systems are metallic , with fermi level lying in partially filled dispersive energy bands . in 1-ml ca@xmath0n ( and sr@xmath0n ) , there are two bands crossing fermi level ( see fig.[fig4]a and [ fig4]d ) . by inspecting the charge density distribution
, one can verify that these two bands are from the 2d confined electron layers residing on the two sides of the [ ca@xmath0n]@xmath1 layer , and were referred to as the 2d electron gas in free space states@xcite .
the two bands are energy splitted due to the coupling between the two sides .
when a second [ ca@xmath0n ] layer is added , i.e. for a 2-ml structure , there appears an additional band crossing fermi level , which has similar dispersion as the two surface bands but is lower in energy ( see fig.[fig4]b ) .
this band is from the confined anionic electron layer between the two [ ca@xmath0n]@xmath1 layers , as we will show in the following .
the energy for this 2d band is lower because the confinement in the interlayer region is stronger than that for the surface , leading to a deeper potential well .
one also observes that the splitting between the two surface bands is reduced , as a result of the reduced coupling between the two with increasing structural thickness .
when going to 3-ml , there are two interlayer gap regions and indeed there is one more band crossing fermi level ( see fig.[fig4]c ) .
one observes that the two bands from the interlayer confined 2d electrons almost coincide , indicating that these states are strongly confined and coupling between neighboring interlayer regions is very small .
this clearly demonstrates the 2d character of the confinement topology .
the above features persist when more [ ca@xmath0n ] layers are added .
then more interlayer confined anionic electron bands will appear and are nearly degenerate in energy . their dispersion and bandwidth
are almost independent on the thickness . for sr@xmath0n ,
the band structure is quite similar to that of ca@xmath0n with the same number of layers ( with a slight decrease of the bandwidth , c.f .
fig.[fig4]d - f ) , implying that the 2d electride character is also maintained in its thin film form . to visualize the real space distribution of the electronic states ,
we take 2-ml ca@xmath0n as an example , and plot their partial electron densities for states within 0.05ev around fermi energy @xmath23 in fig.[fig5]a .
it is clear that the states around fermi level are mainly located in three confined 2d regions : outside the two surfaces and in the interlayer space between the [ ca@xmath0n]@xmath1 layers .
we further plot the total charge densities of each of the three conduction bands for 2-ml ca@xmath0n . as shown in fig.[fig5]b - d ,
the lower band is from the 2d states in between the [ ca@xmath0n]@xmath1 layers , while the two higher bands are from the states confined at the surfaces .
similar result holds for sr@xmath0n .
this analysis validates our claims in the previous discussion .
electron - localization function ( elf ) is useful for the analysis of the degree of electron localization and the bonding character@xcite . in fig.[fig6]a and [ fig6]c
, we show the elf maps of 2-ml ca@xmath0n and 2-ml sr@xmath0n .
three delocalized electron layers can be seen .
one observes that there is no bonding localization attractor between the confined electron layers and the [ ca@xmath0n]@xmath1 ( [ sr@xmath0n]@xmath1 ) layers , indicating the bonding between them is of ionic type .
when one valence electron is removed , the surface confined layers will be vacated while the interlayer anionic electron layer remain largely intact , as shown in fig.[fig6]b and [ fig6]d .
this feature also implies that the out - of - plane work function of these layered materials are mainly determined by the 2d electron layers on the surface .
previous studies found that bulk ca@xmath0n has a highly anisotropic work function ranging from 2.6ev ( for ( 100 ) surface ) to 3.4ev ( for ( 001 ) surface)@xcite . in fig.[fig7
] , we plot our results of the out - of - plane work function as a function of the number of layers for both ca@xmath0n and sr@xmath0n .
one notes that there is a sharp decrease of the work function from 1-ml to 2-ml for both materials , due to the appearance of the strongly confined interlayer states that decouples the two surfaces .
for ca@xmath0n , with increasing layer number , the work function approaches a value around 3.4ev , which is consistent with previous studies .
the work function of sr@xmath0n is less than that of ca@xmath0n by about 0.2ev , showing that the electrons in sr@xmath0n are more loosely bound .
we average the partial electron density around fermi level in the layer plane .
the resulting 1d density profiles for 2-ml with and without one electron removed are plotted in fig.[fig8 ] .
one observes that there are three peaks corresponding to the three anionic electron layers .
after one electron is taken away , the two side peaks disappear while the central peak remains .
the two side peaks of sr@xmath0n have larger width than those of ca@xmath0n , implying a higher degree of delocalization of these surface states hence a lower work function . the first - principles calculation based on dft
has proved to be a powerful tool for the study of dielectric function for metals , including ultrathin metallic films down to few - layer thickness@xcite . in crystalline solids ,
the dielectric function @xmath24 consists of two contributions : a drude - like intraband contribution and an interband contribution .
the imaginary part of the interband contribution involves the interband matrix elements of the momentum operators , and can be evaluated directly in dft@xcite .
its real part can then be calculated via the kramers - kronig relation .
the intraband contribution is typically treated by the drude model , in which the plasma frequency ( tensor ) @xmath25 can be evaluated by dft@xcite .
the calculation details are presented in the method section .
we first consider the dielectric properties of bulk ca@xmath0n and sr@xmath0n .
simple metals such as au , ag , and al generally have a large isotropic bulk plasma frequency around 10ev . in contrast , ca@xmath0n and sr@xmath0n are of layered structure , hence their dielectric properties are expected to exhibit intrinsic anisotropy .
we find that bulk ca@xmath0n has an in - plane ( i.e. in the layer plane ) plasma frequency @xmath26 of 3.14ev and an out - of - plane plasma frequency @xmath27 of 0.95ev ; bulk sr@xmath0n has an in - plane plasma frequency of 2.94ev and an out - of - plane plasma frequency of 1.05ev .
these frequencies are significantly lower than those for simple metals , partly because the conducting electron density is lower . for each material ,
the out - of - plane plasma frequency is much lower compared with the in - plane plasma frequency , showing that the neighboring anionic electron layers are only weakly coupled . in fig.[fig9 ] , we plot the real and imaginary part of in - plane and out - of - plane dielectric functions for the two materials .
they show the character of metallic behavior , i.e. drude peaks at low energy due to intraband contribution and the real part of @xmath24 crossing from negative value to positive value with increasing frequency .
the strong anisotropy in the dielectric function is also quite obvious .
remarkably , because of this anisotropy , the real part of in - plane dielectric function , @xmath28 , changes sign at a frequency which is different from that for @xmath29 , leading to an extended frequency window in which the two components have different signs ( marked as the blue shaded regions in fig.[fig9 ] ) .
this sign difference is the characteristic feature of so - called indefinite media@xcite , which was proposed in the study of metamaterials and has important potential applications such as near - field focusing and building hyperlenses that can transform evanescent fields into propagating modes@xcite .
previous realizations of indefinite media are mostly in artificially assembled structures which require complicated fabrication process and usually have high dissipation .
our results suggest that crystalline solids ca@xmath0n and sr@xmath0n in their bulk form would just be indefinite materials for a frequency range spanning the near infrared .
moreover , one notes that the imaginary parts of @xmath30 and @xmath31 in this frequency range are very small ( @xmath32 ) , implying that they could be ideal low - loss indefinite materials .
next we focus on the dielectric functions of ca@xmath0n and sr@xmath0n few - layer structures . with decreasing film thickness
, one expects that the anisotropy effects would be even stronger .
the intraband contribution to the out - of - plane dielectric response @xmath31 becomes negligible , and it has been shown that for simple metals which are isotropic in bulk can exhibit strongly anisotropic @xmath24 down to few - layer thickness@xcite . in fig.[fig10]a , we plot the results of in - plane plasma frequency @xmath26 as a function of the number of layers for both materials .
one observes that the frequency slowly decreases with increasing number of layers and approaches its bulk value .
the plasma frequency of sr@xmath0n is lower than that of ca@xmath0n by about 0.2ev for each thickness .
the square of plasma frequency is roughly proportional to the product of the carrier density and the inverse of effective mass . in fig.[fig10]b
, one can see that from 1-ml to 2-ml , there is in fact a sharp decrease in the carrier density , as characterized by the dos per unit volume around fermi level , primarily due to the appearance of the large interlayer space of 2-ml .
however , from the band structure in fig.[fig4]b , one notices that the additional 2d electron band ( marked in red color ) has a smaller effective mass around fermi level than the two surface bands , hence decreasing the average effective mass .
this compensates the decrease in carrier density , resulting in an overall small change in the plasma frequency between 1-ml and 2-ml . in fig.[fig11 ]
, we show the interband contribution to the imaginary part of the in - plane dielectric function , @xmath33 , which is closely connected to band structure features .
one noticeable feature is the sharp peak around @xmath34ev for both materials at 1-ml thickness .
this peak has a sharp drop from 1-ml to 2-ml and shifts to a lower energy .
its appearance can be attributed to the ( almost ) parallel sections of the two surface bands for 1-ml , as indicated in the red shaded regions in fig.[fig4]a and [ fig4]d . from 1-ml to 2-ml
, the splitting between two surface bands decreases and the volume increases a lot due to the appearance of the large interlayer region , resulting in the observed change of the peak . for thicknesses larger than 1-ml , there is another peak around 1ev , which is due to the transition between the interlayer 2d bands and the surface bands ( c.f .
fig.[fig4 ] ) . from the band structures ,
it is clear that the response below 1ev is mainly contributed by the states from the 2d anionic electron layers .
the total in - plane and out - of - plane dielectric function with their real and imaginary parts for ca@xmath0n few - layers are plotted in fig.[fig12 ] . for @xmath30 ,
the low energy part is dominated by the intraband drude - like contribution .
@xmath28 is negative below about 1ev .
meanwhile , the intraband contribution is negligible for @xmath35 . hence @xmath29 is largely positive for the low energy part .
the sign difference between @xmath36 and @xmath37 in the low energy range again signals a possible indefinite material .
this mechanism for the anisotropic dielectric function has been discussed in the case of au ultrathin films@xcite .
one observes that the change of @xmath24 against thickness is most dramatic from 1-ml to 2-ml , due to the introduction of the first interlayer anionic electron layer .
the imaginary part of dielectric function for 1-ml is much larger than that for 2-ml , particularly in the near infrared frequency range , which indicates more energy dissipation for possible plasmon modes as we discuss later .
in addition , the variation of @xmath30 versus thickness is overall less dramatic compared with that of @xmath31 , which again reflects the fact that the conducting electrons are strongly confined within the 2d in - plane layered regions . due to the same reason that the different conducting 2d layers are only weakly coupled , the results for both in - plane and out - of - plane components converges rapidly with increasing thickness ( note that the low energy part of the out - of - plane component would eventually be dominated by a drude - like intraband contribution when approaching the bulk limit ) .
one notes that there is only small difference between the results of 4-ml and 5-ml .
we also calculated 8-ml which is the largest system size within our computing capability .
the obtained results are very close to those for 5-ml .
similar features are also demonstrated in the results for sr@xmath0n few - layer structures , as shown in fig.[fig13 ] .
surface plasmon modes are confined electromagnetic excitations propagating at an interface between a conductor and a dielectric@xcite .
it typically requires a sign change of @xmath38 across the interface .
therefore the most commonly used materials for plasmonic applications are metals such as au , ag , and al which have a range of frequencies with negative @xmath39@xcite .
since ca@xmath0n and sr@xmath0n are also metals , one may also wonder whether they could also support surface plasmon modes .
this question is particularly interesting when considering the frequency range in which the material shows indefinite medium property . in such case
, one expects that the positive @xmath40 component may compete with the negative @xmath41 component and tend to destroy the bounded plasmon modes .
let s first consider the interface between bulk ca@xmath0n or sr@xmath0n and a dielectric medium characterized by a frequency - independent dielectric constant @xmath42 .
the interface is parallel to the layer plane .
as we have shown , ca@xmath0n and sr@xmath0n in the bulk form already have strong anisotropy in their dielectric functions .
assume that the interface supports a transverse magnetic ( tm ) plasmon mode travelling along the interface with a wave vector @xmath43 .
following standard derivation using maxwell s equations@xcite , one obtains that @xmath44 here @xmath45 is the wave vector in vacuum , @xmath9 is the speed of light , and @xmath30 and @xmath31 are the two components of the dielectric function for our conducting material . if @xmath46 , the result in eq.([bulkbeta ] ) reduces to the familiar result for isotropic metal@xcite .
the dispersion characteristics of the surface plasmon modes for both materials are shown in fig.[fig14 ] , with @xmath47 ( appropriate for sio@xmath0 ) . in fig.[fig14]a ,
the modes lying to the right of the light line ( in the dielectric medium ) are bounded to the surface .
one observes that the results show characteristic surface plasmon peaks similar to simple metals and confined plasmon modes still exist within the frequency range where the material shows indefinite medium property .
a major difference is that for simple metals the peak position , known as surface plasmon frequency @xmath48 , occurs at higher energies , e.g. @xmath49ev for ag / sio@xmath0 interface@xcite ; while @xmath48 is much lower here , around 1.1ev-1.2ev in the near infrared range .
the @xmath50 shown in fig.[fig14]b is connected to the energy damping during the mode propagation . since high dissipation occurs around @xmath48 , for practical applications , modes with frequencies less than @xmath48
are used . here for ca@xmath0n , if we take @xmath51ev , the corresponding surface plasmon wavelength @xmath52 nm , and we have a long propagation length @xmath53 m and the decay length in the dielectric is @xmath54 nm . longer propagation length can be achieved at lower frequencies , e.g. at @xmath55ev , we can have @xmath56 m but the mode confinement is reduced , with @xmath57 nm .
these values are comparable to those for the noble metals ( in the visible or uv frequency range ) which are the usual building blocks for plasmonics devices .
however , the operating frequency here is much lower . from above discussion
, we see that despite the intrinsic anisotropy , bulk ca@xmath0n and sr@xmath0n can still support surface plasmons and could be suitable plasmonics materials in the near infrared frequency range .
we then turn to the thin films of ca@xmath0n and sr@xmath0n sandwiched between dielectric materials .
besides the change in the thickness - dependent dielectric function , an important effect in thin films is that the surface plasmon modes at two interfaces could couple and form two modes with opposite parity : a symmetric mode ( @xmath58 ) and an antisymmetric mode ( @xmath59)@xcite . their dispersions have been derived before in the study of ultrathin metallic films@xcite , and are quoted in our method section . in fig.[fig15 ] , we plot the surface plasmon dispersion characteristics for ca@xmath0n with different film thicknesses .
one observes that the antisymmetric modes are lying on the light line of the dielectric medium , indicating that they are squeezed out of the metal region forming unbounded modes propagating in the dielectric material , which is similar to the case of au ultrathin films
. meanwhile , pronounced plasmonic peaks do show up for the symmetric modes , clearly indicating that these are bounded surface plasmon modes .
again , one observes that the variation with thickness is most dramatic between 1-ml and 2-ml . for 1-ml ,
the bounded modes occur around 1ev .
the high peak in @xmath60 vs. @xmath61 shows the modes are strongly bounded to the metallic layer . for thicknesses of 2-ml to 5-ml ,
the dispersions are quite close .
the corresponding surface plasmon frequencies shift to around 1.2ev . as for the imaginary part of @xmath43
, 1-ml structure has relatively large values , whereas @xmath50 for 2-ml to 5-ml almost collapse on a single curve and deceases rapidly in the range below 0.8ev .
for sr@xmath0n few - layers , the results for surface plasmon dispersion show similar features . the dispersions for symmetric ( @xmath58 ) modes are shown in fig.[fig16 ] .
the antisymmetric ( @xmath59 ) modes are again unbounded hence are not shown .
compared with their bulk results , one observes that the surface plasmon frequency @xmath48 is more or less the same , but both @xmath60 and @xmath50 are greatly increased by two orders of magnitude .
this means that the plasmon wavelength and the confinement scale are much smaller , which are desired .
however , the propagation length is also decreased at the same time .
therefore , it is more meaningful to consider the dimensionless ratio @xmath62 , which measures how many surface plasmon wavelengths can be covered before the wave loses most of its energy , as well as the wave localization ( or wave shrinkage ) quantified by @xmath63 , where @xmath64 is the wavelength in air . in fig.[fig17 ] , we show these two dimensionless characteristics for the two materials as functions of @xmath65 .
the results for 3-ml and 4-ml are similar to those of 5-ml hence are not shown .
one observes that for 1-ml @xmath62 is lower for most frequencies , due to the relatively high dissipation associated with @xmath50 ( c.f .
fig.[fig15]d ) . generally , the wave localization reaches its peak near the surface plasmon resonance .
however , there @xmath62 is small due to the enhanced dissipation .
while larger @xmath62 can be achieved for longer wavelengths , the wave localization becomes poor in that range .
this tradeoff is a typical feature for surface plasmons . for application purposes
, a compromise can be reached somewhere in between when the two are comparable . since 1-ml ca@xmath0n and sr@xmath0n have considerably larger dissipation , in the following
we take 2-ml ca@xmath0n and sr@xmath0n as examples . for @xmath66 m
( 0.64ev ) in the near infrared range , 2-ml ca@xmath0n has a decay length @xmath67 nm in the dielectric , showing that the plasmon mode is strongly confined to the metallic layer .
the ratios @xmath68 and @xmath69 .
these two values are considerably larger than what noble metals could achieve at their optimal working frequencies .
the corresponding values for sr@xmath0n at @xmath66 m ( @xmath55ev ) are given by @xmath70 nm , @xmath71 and @xmath72 . in comparison , the surface plasmon modes for noble metals such as au and ag are unbounded at this wavelength .
noticeably , the surface plasmon wavelengths in these materials are very small , @xmath73 nm for ca@xmath0n and @xmath74 nm for sr@xmath0n ( at @xmath55ev ) , while the shortest @xmath75 for au and ag ( occur in the visible light range ) would be at least larger than 100 nm .
these suggest that ca@xmath0n and sr@xmath0n thin films have great potential for making plasmonic devices operating in the near infrared range with much smaller scales .
the loosely bound anionic electrons in electrides are highly reactive in ambient conditions
. they could be used as good electron donors and as excellent catalysts for chemical reactions@xcite . however , for physical applications
, one needs to stabilize its property , e.g. by effective encapsulation . for 1-ml ca@xmath0n ,
a possible encapsulation scheme using 2d insulating layers of graphane was proposed@xcite . in the case of few - layer alkaline earth subnitrides , since as we discussed each interlayer anionic electron layer are strongly confined in 2d regions and coupling between layers is small , one expects that the the most reactive electrons are from the surface layers , while the electron layers inside should be less reactive .
a detailed study of this point and possible encapsulation schemes for sr@xmath0n will be deferred to a future work . compared with ca@xmath0n
, sr@xmath0n has a smaller electrostatic potential associated with the larger atomic number of sr .
this was reflected in the lower work function , and may also lead to a high electron mobility for sr@xmath0n few - layers@xcite .
previous experimental studies have shown that bulk ca@xmath0n has high mobility of 520cm@xmath19/(v@xmath77s)@xcite .
one expects that the mobility for sr@xmath0n may be even higher .
hence sr@xmath0n few - layers as new 2d conducting materials could have good potential for electronics applications . as having been demonstrated in other 2d layered materials , strain engineering has proved to be a powerful tool to modify and control the material properties .
the layered materials discussed here have a large interlayer spacing ( @xmath10 ) between the [ x@xmath0n]@xmath1 layers .
we also expect that applying strain could be a good method to tune the material properties such as the dielectric function , the plasmonic dispersion , and the carrier mobility .
a systematic study of the strain effects is currently underway .
first - principles calculations were carried out using the vienna _ ab_-initio simulation package ( vasp)@xcite , based on the density functional theory ( dft ) .
the exchange - correlation functional was treated using perdew - burke - ernzerhof generalized gradient approximation@xcite .
the projector augmented wave ( paw ) pseudopotential method@xcite is employed to model interactions between electrons and ions .
the cutoff for plane - wave expansion is set to be 600 ev .
the vertical distance between thin films ( the thickness of the vacuum gap ) is at least 18 , which is large enough to avoid artificial interactions between the film and its periodic images .
both the atomic positions and lattice constant were fully relaxed using conjugate gradient method .
the convergence criteria for energy and force were set to be 10@xmath78ev and 0.01ev/ , respectively .
dft - d2 method was applied to describe the long - range van der waals interaction .
the brillouin zone integrations have been carried out on a @xmath17-centered @xmath79-mesh .
monkhorst - pack k - point meshes@xcite with sizes of 15@xmath8015@xmath801 and 31@xmath8031@xmath801 were used for geometry optimization and static electronic structure calculation , respectively . in the later study of optical properties ,
the sizes of @xmath79-mesh are significantly increased to 61@xmath8061@xmath801 and 41@xmath8041@xmath809 for thin films and bulk respectively , to achieve highly converged results . for the integration over the brillouin zone in calculating dielectric functions
, we used the first order methfessel - paxton method@xcite with a value of 0.1ev .
the phonon dispersions of the structures were calculated by using density functional perturbation theory as implemented in phonopy code@xcite .
the optical properties of solids are mainly due to the response of the electron system to a time - dependent electromagnetic perturbation . for metals ,
the optical complex dielectric function consists of interband and drude - like intraband contributions : @xmath81 the imaginary part of the interband part can be calculated using the results from dft calculations as@xcite @xmath82&=&\frac{4\pi^2e^2}{v}\lim_{q\rightarrow 0}\frac{1}{q^2}\sum_{nm;\bm k } 2f_{n\bm k}\langle u_{m,\bm k+q\bm e_\alpha}|u_{n,\bm k}\rangle \langle u_{n,\bm k}|u_{m,\bm k+q\bm e_\beta}\rangle\cr & & \qquad\qquad \times\left[\delta(e_{m,\bm k}-e_{n,\bm
k}-\omega)-\delta(e_{m,\bm k}-e_{n,\bm k}+\omega)\frac{}{}\right],\end{aligned}\ ] ] where @xmath83 and @xmath43 refer to cartesian coordinates , @xmath84 are unit vectors , @xmath85 is the volume of the unit cell , @xmath86 and @xmath87 are the periodic part of the bloch wave function and the corresponding eigenenergy for band @xmath88 and wave vector @xmath89 , and @xmath90 is the fermi - dirac distribution function . the real part of interband contribution can be obtained through the kramers - kronig relation .
the intraband contribution is usually modeled by the drude model : @xmath91=\frac{\gamma\omega^2_{p,\alpha\beta}}{\omega(\omega^2+\gamma^2)},\ ] ] @xmath92=1-\frac{\omega^2_{p,\alpha\beta}}{\omega^2+\gamma^2}.\ ] ] here @xmath93 is a life - time broadening obtained either from a higher - order calculation or from experiments . in our calculation
, we used the experimental determined electron life - time ( of 0.6ps)@xcite for bulk ca@xmath0n to estimate @xmath93 ( @xmath94mev ) .
the same value was also used for the calculations of sr@xmath0n due to their similar electronic structures .
we have checked the sensitivity of our results dependence on @xmath93 by repeating the calculations of @xmath24 with @xmath93 value varying from @xmath95 to @xmath96 .
the obtained results are of little difference .
this is because the value of @xmath93 is already quite small ( reflecting the fact that these materials are good metals ) .
the @xmath97 is the plasma frequency tensor which can be calculated using @xmath98 the dielectric functions and plasma frequencies were suitably renormalized to exclude the vacuum region from the unit cell in our calculations . for thin film structures , the plasmon modes at two surfaces would couple and form two modes with different parity@xcite .
the equations governing their dispersions have been derived before@xcite : @xmath99 @xmath100 the first equation above is for the antisymmetric ( @xmath59 ) mode and the second equation is for the symmetric ( @xmath58 ) mode . here
@xmath101 in the equations is the film thickness .
other quantities in these two equations are defined in the main text . in our calculation
, we solve the two equations numerically using a two - dimensional unconstrained nelder - mead minimization algorithm@xcite with a tolerance of @xmath102nm@xmath103 in the complex wave vectors .
toda , y. , yanagi , h. , ikenaga , e. , kim , j. j. , kobata , m. , ueda , s. , kamiya , t. , hirano , m. , kobayashi , k. & hosono , h. work function of a room - temperature , stable electride [ ca@xmath108al@xmath109o@xmath110@xmath107(e@xmath5)@xmath111 .
_ * 19 * , 3564 - 3569 ( 2007 ) .
fang , c. m. , de wijs , g. a. , de groot , r. a. , hintzen , h. t. & de with , g. bulk and surface electronic structure of the layered sub - nitrides ca@xmath0n and sr@xmath0n , _ chem . mater .
_ * 12 * , 1847 - 1852 ( 2000 ) .
kitano , m. , inoue , y. , yamazaki , y. , hayashi , f. , kanbara , s. , matsuishi , s. , yokoyama , t. , kim , s. w. , hara , m. & hosono , h. ammonia synthesis using a stable electride as an electron donor and reversible hydrogen store .
_ nat chem _ * 4 * , 934 - 940 ( 2012 ) .
zhao , s. t. , li , z. y. & yang , j. l. obtaining two - dimensional electron gas in free space without resorting to electron doping : an electride based design .
* 136 * , 13313 - 13318 ( 2014 ) .
gregory , d. h. , bowman , a. , baker , c. f. & weston , d. p. dicalcium nitride , ca@xmath0n - a 2d @xmath113excess electron@xmath113 compound : synthetic routes and crystal chemistry .
_ j. mater .
chem . _ * 10 * , 1635 ( 2000 ) .
laref , s. , cao , j. r. , asaduzzaman , a. , runge , k. , deymier , p. , ziolkowski , r. w. , miyawaki , m. & muralidharan , k. size - dependent permittivity and intrinsic optical anisotropy of nanometric gold thin films : a density functional theory study .
express _ * 21 * , 11827 - 11838 ( 2013 ) .
dionne , j. a. , sweatlock , l. a. , atwater , h. a. & polman , a. planar metal plasmon waveguides : frequency - dependent dispersion , propagation , localization , and loss beyond the free electron model .
b _ * 72 * , 075405 ( 2005 ) .
kresse , g. & hafner , j. _ ab initio _ molecular dynamics for open - shell transition metals .
b _ * 48 * , 13115 - 13118 ( 1993 ) .
kresse , g. & furthmller , j. efficient iterative schemes for _ ab initio _ total - energy calculations using a plane - wave basis set .
b _ * 54 * , 11169 - 11186 ( 1996 ) .
gonze , x. & lee , c. y. dynamical matrices , born effective charges , dielectric permittivity tensors , and interatomic force constants from density - functional perturbation theory .
b _ * 55 * , 10355 - 10368 ( 1997 ) .
the authors thank d.l .
deng for helpful discussions .
this work was supported by the most project of china ( nos .
2014cb920903 , 2011cba00100 ) , the nsf of china ( nos . 11174337 , 11225418 ) , the srfdphe of china ( no . 20121101110046 ) and the sutd - srg - epd2013062 . y.y .
and s.a.y .
conceived the idea and supervised the work .
s.g . performed the calculation and the data analysis .
, s.a.y . , and y.y .
contributed to the interpretation of the results and wrote the manuscript .
all authors contributed in the discussion and reviewed the manuscript .
cccccccc + & & & & & & & + & + 1 & 3.562 & 2.516 & 2.516 & & & & + 2 & 3.562 & 8.608 & 2.509 & & & 3.598 & + 3 & 3.552 & 14.957 & 2.514 & 2.504 & & 3.712 & + 4 & 3.550 & 20.099 & 2.516 & 2.507 & & 3.721 & 3.793 + 5 & 3.555 & 26.419 & 2.513 & 2.505 & 2.508 & 3.703 & 3.823 + & + 1 & 3.778 & 2.761 & 2.761 & & & & + 2 & 3.775 & 9.427 & 2.757 & & & 3.913 & + 3 & 3.770 & 16.260 & 2.760 & 2.748 & & 3.983 & + 4 & 3.765 & 21.821 & 2.764 & 2.754 & & 4.016 & 4.068 + 5 & 3.768 & 28.562 & 2.758 & 2.751 & 2.753 & 3.989 & 4.061 + | based on first - principles calculations , we systematically study the electronic , dielectric , and plasmonic properties of two - dimensional ( 2d ) electride materials x@xmath0n ( x = ca , sr ) . we show that both ca@xmath0n and sr@xmath0n are stable down to monolayer thickness . for thicknesses larger than 1-monolayer ( 1-ml ) ,
there are 2d anionic electron layers confined in the regions between the [ x@xmath0n]@xmath1 layers .
these electron layers are strongly trapped and have weak coupling between each other . as a result , for the thickness dependence of many properties such as the surface energy , work function , and dielectric function , the most dramatic change occurs when going from 1-ml to 2-ml . for both bulk and few - layer ca@xmath0n and sr@xmath0n ,
the in - plane and out - of - plane real components of their dielectric functions have different signs in an extended frequency range covering the near infrared , indicating their potential applications as indefinite media .
we find that bulk ca@xmath0n and sr@xmath0n could support surface plasmon modes in the near infrared range . moreover
, tightly - bounded plasmon modes could exist in their few - layer structures .
these modes have significantly shorter wavelengths ( few tens of nanometers ) compared with that of conventional noble metal materials , suggesting their great potential for plasmonic devices with much smaller dimensions .
school of physics , beijing institute of technology , beijing 100081 , china research laboratory for quantum materials and epd pillar , singapore university of technology and design , singapore 487372 , singapore school of physics and electronic & electrical engineering , huaiyin normal university , huaian 223300 , china @xmath2e - mail : [email protected] ; [email protected] electrides are a special kind of ionic solids with cavity - trapped electrons serving as the anions@xcite .
these electrons are spatially separated from the cations in a regular crystalline array , and are not bound to any particular atom , molecule , or bond .
the first crystalline electride , cs@xmath3(18-crown-6)@xmath4e@xmath5 , was synthesized in 1983@xcite , and several other electrides have been successfully discovered or predicted later on@xcite . the properties of electrides are closely connected to the topology and geometry of the cavities which confine the anionic electrons@xcite .
the early examples of electride demonstrate confinement of zero - dimensional cavities or one - dimensional weakly linked channels@xcite .
recently , a new type of electride with two - dimensional ( 2d ) confinement of anionic electrons was discovered in dicalcium nitride ( ca@xmath0n ) which has a layered structure@xcite . from formal valence consideration , each unit cell of ca@xmath0n
should have one excess electron .
experimental measurements on properties such as the transport parameters and work functions , combined with first - principles calculations , indeed proved that there is a built - in anionic electron layer confined between the calcium layers , which agrees well with the chemical formula [ ca@xmath0n]@xmath3e@xmath5@xcite . this discovery of 2d electride materials ( here 2d refers to the anionic electron confinement topology ) has generated great research interest .
first - principles calculations have predicted several other 2d electrides@xcite , including the other two alkaline earth sub - nitrides sr@xmath0n and ba@xmath0n@xcite .
recent theoretical work also indicated that the monolayer ca@xmath0n might be mechanically exfoliated from the bulk , meanwhile , its electron confined layer could be maintained , and suitable encapsulation layers were proposed to protect it in ambient environment@xcite .
motivated by these recent experimental and theoretical progress on 2d electrides , and also by the surge of research activities on 2d materials in recent years initiated by the discovery of graphene , in this work , we conduct a systematically investigation of the electronic , dielectric , and plasmonic properties of monolayer and few - layer alkaline earth sub - nitrides ca@xmath0n and sr@xmath0n .
we find that besides ca@xmath0n , sr@xmath0n is also dynamically stable down to monolayer thickness .
their phonon spectra exhibit characteristic features of 2d materials . for thickness
greater than 1-ml , besides the two surface electron bands , there are additional bands crossing the fermi energy which are due to the anionic electron layers confined in the 2d interlayer regions between the [ x@xmath0n]@xmath1 layers .
these interlayer bands are lower in energy than the surface bands and are nearly degenerate , indicating that they are only weakly coupled .
the thickness dependence of properties such as surface energy , work function , and dielectric function are analyzed .
the change in property with thickness is most dramatic when going from 1-ml to 2-ml , which is associated with the appearance of the first interlayer electron band .
we find that sr@xmath0n has a lower surface energy and lower work function compared with ca@xmath0n . due to their intrinsic structural anisotropy ,
these layered materials have highly anisotropic dielectric functions .
in particular , the in - plane and out - of - plane components of dielectric function can have different signs , which occurs already for the bulk form in the near infrared frequency range .
moreover the dissipation is low in the range for the bulk materials .
this shows that bulk ca@xmath0n and sr@xmath0n could be ideal low - loss indefinite media@xcite .
we further investigate the possibility of guided surface plasmon modes for these metallic materials and show that bulk ca@xmath0n and sr@xmath0n have good performance comparable to the noble metals but operating at lower frequencies , while their thin film structure can have strongly bounded plasmon modes which have advantage of much shorter wavelength compared to the conventional noble metals .
our findings thus identify a class of promising 2d electride materials and reveal their potentials for future electronics and plasmonics applications . |
at the present time , the greatest mystery posed by microlensing is the cause of the lensing events towards the large magellanic cloud ( lmc ) .
one of the original motivations for microlensing experiments was the search for halo dark matter constituents such as brown dwarfs .
based on the 6 - 8 events recorded during the first two years of their search , the macho group determined that if these events were caused by an intervening halo lensing population , this population would account for a good fraction of the mass of the halo .
however , they also concluded that the mass of the lenses was greater than 0.1 @xmath3 , ruling out a brown dwarf candidate ( alcock 1997a ) .
white dwarfs and other stellar remnants have been eliminated by a closed box model of halo chemical evolution ( gibson & mould 1997 ) and through a more general model of cosmological chemical evolution ( fields , freese & graff 1998 ) .
currently , , there is no really viable candidate for a halo lensing population for the lmc .
a great limitation of using the lmc to probe the galactic halo for microlensing events is that it only samples a single line of sight . in response
, there have been attempts to look for halo microlensing along two other lines of sight , towards m31 ( crotts & tomaney 1996 ; ansari 1997 ) and towards the small magellanic cloud ( smc ) .
indeed , two events have so far been detected towards the smc .
both of these events proved to be different from the ensemble of other events observed towards the lmc .
the first event , macho-97-smc-1 ( alcock 1997c , palanque - delabrouille 1998 ) had a time scale of 120 days , much longer than those of the events seen towards the lmc .
an event of such long duration should show a measurable parallax deviation from the standard light curve due to the circular motion of the earth around the sun , unless the mass of the lens is very large or the lens is in the smc .
the lack of a parallax deviation enabled the eros group to show that this event is likely due to a lens within the smc , although it may be due to a high mass lens in the halo .
( palanque - delabrouille 1998 ) .
the long time scale of this event also suggested that the lens was drawn from a different distribution from that of the lmc lenses .
the second observed smc event , macho-98-smc-1 , was also unique .
it was alerted by the macho group to be a caustic crossing event caused by a binary lens ( becker 1998 , alcock 1998 ) .
a caustic crossing allows an actual resolution of the spatial structure of the source star .
detailed observations every few minutes by the eros collaboration ( which observed the end of the caustic crossing ) ( afonso 1998 ) and by the planet collaboration ( which observed the peak of the caustic crossing ) ( albrow 1999 ) and further photometry by macho / gman ( alcock 1998 ) and by ogle ( udalski 1998 ) allowed all the collaborations to place limits or solve for the proper motion of the lens . for this paper , we will adopt planet model 1 with a proper motion of the lens with respect to the source of 1.26 km s@xmath4 kpc@xmath4 , which gives the best fit to all the data ( a. gould , private communication ) .
this proper motion is ( as we will confirm ) consistent with the lens being in the smc ; it is an order of magnitude smaller than the proper motion expected for a lens in the galactic halo . to sum up ,
while the smc has been tapped as an alternative probe of the galactic halo , both of the microlensing events observed so far towards the smc are most likely due to lenses inside the smc .
there have been no precise theoretical predictions of the microlensing statistics of the smc because its structure is poorly understood .
the smc is known to be far from virial equilibrium ( staveley - smith 1997 ) , having suffered from a recent encounter with the lmc as well as tidal stress from the milky way ( murai & fujimoto 1980 ; gardiner , sawa & fujimoto 1994 ) .
one can measure the stellar density in the plane of the sky , but there is no accepted three dimensional model for the distribution of stars in the smc ( westerlund 1997 ) .
mathewson , ford & visvanathan ( 1986 ) examined cepheids in the smc and found a total depth of 32 kpc .
caldwell & coulson ( 1986 ) find a total depth for cepheids in the smc of 20 kpc , based on an inclined plane model .
however , for microlensing , what is important is the line of sight dispersion in distance about that plane : caldwell & coulson ( 1986 ) find a dispersion about the plane of @xmath5 mag .
welch ( 1987 ) find an rms thickness of 0.12 mag corresponding to a total width @xmath6 of 13 kpc .
welch ( 1987 ) suggest that the mathewson , ford and visvanathan ( 1986 ) data suffered from additional unaccounted scatter in part because of poor phase coverage in their sample : many of the cepheids were only observed during luminosity minima . since so little is known about the line of sight structure and transverse velocity distribution of the smc , previous attempts to determine the microlensing statistics have relied on simple models of the smc .
palanque - delabrouille ( 1998 ) assumed a double exponential smc and considered a range of line of sight scale lengths to derive an optical depth of @xmath7 .
sahu & sahu ( 1998 ) examined a few simple models of the smc , with a total line of sight width of @xmath8kpc to derive an optical depth of @xmath9 . since there is no measurement of the transverse velocity distribution in the smc
, palanque - delabrouille ( 1998 ) , albrow ( 1999 ) and sahu & sahu ( 1998 ) all assumed that the transverse velocity distribution should approximately resemble the line of sight distribution .
however , as the smc is elongated along the line of sight , and possibly consists of several distinct components , this can at best be a rough estimate .
as these models are not accurate in detail , or even in their basic parameters , they could only provide an order of magnitude estimate of the optical depth and time scales of microlensing events .
since there are only two detected events , no greater precision was needed to be consistent with the data .
however , several more events would be detected by existing microlensing programs if the present event rate of one per year were to continue , and many more could in principle be detected ( gould 1999 ) .
there is still a need to make a more accurate prediction of the microlensing optical depth due to self - lensing in the smc .
both the line of sight dispersion and the transverse velocity dispersion of the smc will strongly affect the number and character of the microlensing events due to self - lensing in the smc .
the optical depth , the probability that a particular star is being lensed , is directly proportional to the line of sight depth of the smc , while the time scales of the smc events are in part determined by the transverse velocity dispersions of the smc . by studying an accurate model of the smc , we can theoretically predict these quantities . conversely , comparison of the microlensing statistics generated by a model with the data will provide additional information to constrain the structure of the smc . in several respects ,
the most realistic models of the smc are the result of computer simulations .
such models can provide complete information on the 3-d structure and kinematics , which is difficult and in some cases impossible to obtain by other means . in this paper , we examine the _
n-_body simulation by gardiner & noguchi ( 1996 ) ( gn ) .
this simulation represents , to date , the only published detailed self - gravitating simulation of the smc conducted with _ n-_body techniques ; previous work ( e.g. , murai & fujimoto 1980 ) relied exclusively on test particle techniques .
this simulation successfully reproduces a central bar of the smc due to a _ spontaneous _ bar instability as well as other features generated as a result of the tidal interaction of the smc with the lmc ( represented by a fixed plummer potential ) and the galaxy ( represented by a fixed logarithmic halo potential ) .
such tidal features included a wing of the smc , an inter - cloud bridge ( both formed as a result of an smc - lmc encounter @xmath1 yr ago ) and the magellanic stream ( originating at the penultimate perigalactic passage of the magellanic clouds 1.5 gyr ago ) .
regarding the internal structure of the smc , the model produces large - scale features , namely a central bar and tidally induced spiral arms , which appear to be related to corresponding structures delineated by observations of cepheids ( caldwell & coulson 1986 ) .
the kinematics of the model correspond to the pattern seen in the velocity
right ascension plane delineated by hi and young stellar objects found by mathewson et al .
( 1988 ) , showing conclusively that this pattern is due to a bar structure viewed nearly end - on . on the basis of the successful reproduction of many of the observed characteristics of the smc
, we will conduct a detailed microlensing analysis of the simulation by gn . in section 2
, we will examine this simulation to derive a map of the microlensing optical depth of the smc from self - lensing , i.e. , lensing of stars in the back of the smc by stars in the front of the smc .
we will show that the optical depth is concentrated toward the centre of the bar , though some events should still be present towards the wing .
we will show that the gn simulation has a lower optical depth than the most likely value found by the eros and macho experiments ( palanque - delabrouille 1998 , alcock 1998 ) .
however , with only one event in the eros analysis and two in the macho analysis , our low optical depth is still consistent with the experiments at the 90% confidence level . in section 3 , we will calculate the transverse velocities of the lenses with respect to the moving line of sight and use these velocities to calculate proper motions and , assuming a typical lens mass function , the time scales of the lensing events .
the time scales and proper motion distribution in the simulation will be shown to be consistent with those observed .
the gn simulation also follows debris from the smc thrown off by the interaction with the lmc and the galaxy . in section 4
, we will show that this debris does not explain the observed microlensing optical depth towards the lmc .
we conclude by presenting a discussion and summary in sections 5 and 6 .
our calculations of the microlensing statistics of the smc are based on the `` present day '' snapshot of the gn simulation .
the simulation is set up so that the smc is initially represented as a disc galaxy with a halo of equal mass to the disc , and the simulation contains both `` disc '' and `` halo '' particles . at the present epoch ,
the disc and halo particles are somewhat mixed in spatial distribution , and the labels refer only to the initial positions of the particles . under the assumption that the disc particles are stars and
the halo particles are nonlensing dark matter , we define our principal model to be one in which only the disc particles participate in lensing .
we also tested a model in which both halo and disc particles participate in lensing , but with one - half of the mass represented by both types of particles considered to be in non - lensing gas and dark matter .
the two models gave qualitatively similar results .
we calculate the microlensing statistics of our _ n-_body simulation by generating a statistical ensemble of microlensing `` events '' .
a microlensing event is defined to occur when a lensing object lies within an einstein radius of the line of sight to a background `` source '' star .
this probability is generally quite low , which is why microlensing searches require observations of tens of millions of source stars to find a handful of events . to boost our statistics
, we artificially increase the einstein radius of our simulation by a boost factor @xmath10 which increases the probability that a star will be lensed by a factor @xmath11 .
for each source particle in the simulation , we count how many lens particles lie within the boosted einstein radius .
the microlensing optical depth to a particular source particle is then just the number of lenses in front of that particle divided by the boost factor , @xmath12 .
all the microlensing statistics discussed below were extracted from this ensemble of events . a map of the optical depth of the smc calculated for the simulation can be found in fig .
[ tausmcfig ] . the optical depth in the centre of
the smc is relatively high , with a peak of @xmath13 , of the same order as that measured by the microlensing experiments . as can be readily seen in fig .
[ tausmcfig ] , the optical depth of the smc is not constant : it is highest in the centre of the bar and falls off towards the wings .
comparison with the projected mass surface density of our simulation shown in fig .
[ nmapfig ] shows that in the centre of the bar , the optical depth is roughly proportional to the projected mass density . towards the edges of the bar ,
there are some regions with a relatively high optical depth where there is low mass density ; these regions are due to the `` wing '' of material thrown tens of kpc behind the smc by the most recent collision with the lmc .
isolated peaks in the optical depth in this region are due to individual particles in the wing lying well behind the smc .
particle distribution plots showing the wing and bar can be found in gn .
however , the microlensing groups do not measure the peak optical depth of the smc .
they report a single number which is an average optical depth over that portion of the smc that is observed in their fields .
in addition , different regions are weighted according to the local density of source stars in the reference images of the microlensing surveys , a complex function of the surface brightness , exposure time , and seeing of the individual experiments
. the experimental optical depth will thus be @xmath14 where @xmath15 is the optical depth of star @xmath16 and @xmath17 is the total number of resolved stars in the experiment .
there is an additional complication due to blending , which is more severe in the inner regions than the outer ones ; however , monte - carlo simulations suggest that this should be at most a 20% effect ( palanque - delabrouille 1997 ) . the optical depth parameter was originally designed to measure the rate of microlensing due to a foreground galactic halo population , and was thought to be a slowly changing function of position . in that case
, the derived optical depth would be relatively independent of the averaging , and all experiments should arrive at the same number .
however , as can be seen in fig.[tausmcfig ] , this is not the case : the smc self - lensing optical depth rises rapidly to its peak value roughly as fast as the luminosity density of the smc rises .
thus , it is possible that different experiments averaging over different fields will report discrepant values of the optical depth .
for example , the macho experiment has equal exposure times for all its fields . in comparison ,
the eros experiment covers a wider projected area of the smc , and has longer exposures and larger numbers of resolved stars in the outer fields .
thus , the eros experiment places more weight on the outer fields with low optical depth than the macho experiment .
we expect the eros experiment to report a lower average optical depth for smc self - lensing than the macho experiment . as an example of an experimental optical depth calculation from the gn simulation
, we will determine the mean optical depth that would be reported by the eros experiment with which one of us ( dg ) has worked .
the source densities in the eros fields can be found in palanque - delabrouille ( 1997 ) .
we reproduce the number of sources in each ccd in the eros fields in fig.([sourcefig ] ) . we then approximate equation ( [ tauexpt ] ) as an average over the individual ccds in the eros ccd array : @xmath18 to get @xmath19 .
if we assume that the density of source stars is equal to the density of lens stars , we get @xmath20 .
the proper motions of our ensemble of smc events are shown in fig.[mufig ] .
the typical event in the simulation has a proper motion of @xmath21 km / sec / kpc .
also shown is the proper motion of the second smc event . as can be seen from this figure ,
the simulation is easily consistent with the solution for the sole event with an accurately measured proper motion under the assumption that the lens is in the smc .
we calculated the distribution of microlensing time scales ( einstein radius crossing times ) by assuming that the lenses have a flat mass function , @xmath22const between the cut off masses of @xmath23 and @xmath24 .
in this distribution , the average mass of a lens is 0.35@xmath3 .
we find that the mean einstein radius crossing time is 100 days while the median einstein radius crossing time is 78 days ( see fig.[thatfig ] ) .
of course , the mass function of the ordinary stellar population of the smc is very poorly known , especially at the lower end .
we tried an alternative mass function , @xmath25 = @xmath26 between the cutoff masses of 0.01@xmath3 and 1@xmath3 .
note that half of the stellar mass due to this mass function is in brown dwarfs . for this model ,
the mean einstein radius crossing time is 50 days while the median time scale is 30 days .
the long time scales measured in the smc suggest that the mass function of the smc is weighted towards stars of mass 0.3@xmath3 , and is not dominated by brown dwarfs .
zhao ( 1998 ) has proposed that debris lying in a tidal tail stripped from an ancient magellanic progenitor galaxy by the milky way may explain the observed microlensing rate towards the lmc . within this general framework
, he suggests that the debris thrown off by the smc - lmc tidal interaction could also lead to a high optical depth for the lmc .
there have been several observational attempts to search for this debris .
zaritsky & lin ( 1997 ) report a possible detection of such debris in observations of red clump stars , but the results of further variable star searches by the macho group ( alcock 1997b ) , and examination of the surface brightness contours of the lmc ( gould 1998 ) showed that there is no evidence for such a population .
a stellar evolutionary explanation for the observations of zaritsky & lin ( 1997 ) was proposed by beaulieu & sackett ( 1998 ) . however , possible evidence for debris within a few kpc of the lmc along the line of sight is reported by the eros group ( graff in preparation ) . in the gn simulation , there is some debris behind the lmc , but not enough to explain the observed microlensing events .
the average optical depth of the lmc due to this background debris is only of order a few@xmath27 , two orders of magnitude below that observed . even though the optical depth of the individual stars in the background debris may be high due to the foreground lmc
, the observed optical depth will be low since the vast majority of observed stars will be in the foreground lmc , not in the background debris . nor is this a matter of bad luck in the placement of the lmc with respect to the distribution of lmc debris in the simulation .
nowhere outside the smc does the optical depth due to smc debris rise above a @xmath28few@xmath2 , orders of magnitude below that observed by the macho group .
it seems unlikely , based on the gn simulation , that the smc could have produced enough debris to give rise to a high optical depth for the lmc .
the optical depth that we derive for the smc , @xmath29 , is much lower than the optical depths derived in the simple models of palanque - delabrouille ( 1998 ) of @xmath7 and sahu & sahu ( 1998 ) of @xmath30 .
we will examine their assumptions and show why their optical depths are larger than ours .
the discrepancy arises because the smc in the gn simulation has a lower line of sight distance dispersion than was assumed by palanque - delabrouille ( 1998 ) and by sahu & sahu ( 1998 ) .
palanque - delabrouille ( 1998 ) assume that the density of source stars follows the density of light whereas in fact it is actually roughly constant .
this is not a bad approximation however , and they only overestimate the optical depth by a factor of 1.3 .
we caution that if the eros fields had been larger , the discrepancy would have been greater .
sahu & sahu ( 1998 ) derive an average mass density by dividing a fiducial value of the total mass of the smc of @xmath31 by a fiducial total area of 15 kpc@xmath32 to obtain a mass density of @xmath33pc@xmath32 .
the use of an average mass density is appropriate in the limit that the source density is constant .
this is roughly true for the eros fields which had longer exposures for the outer fields ( see fig.[sourcefig ] ) , but is likely to be false for the macho fields which had constant exposures .
in fact , given that the eros source stars have roughly unform density , the average mass density should be computed by dividing the mass within the eros fields by the area of those fields .
the eros fields cover an area of about 10 kpc@xmath32 and a total mass of about @xmath34 ( palanque - delabrouille 1998 ) so these estimates are not far off , yielding a total surface density of @xmath35pc@xmath32 , close to the values derived by sahu & sahu ( 1998 ) .
we find that the core of the smc bar in the gn simulation , where most of the microlensing takes place in the simulation , is actually rather narrow , with a line of sight rms dispersion of only 1 kpc ( see fig.[pencilfig ] ) .
sahu & sahu ( 1998 ) and palanque - delabrouille ( 1998 ) assume larger dispersions based on the cepheid observations of caldwell & coulson ( 1986 ) and mathewson , ford & visvanathan ( 1986 ) .
the discrepancy between the simulation and the cepheid observations was discussed by gn who suggested that , being a young population , cepheids may be influenced by gas - dynamics not present in the strictly _
n-_body simulation .
however , since there is no reason to believe that the lenses in the smc are due to a particularly young population , they may in fact be better represented by the _
n-_body particles of the simulation than by the ( young ) cepheid population .
however , measuring the thickness of a cepheid population is quite difficult ( welch 1987 ) .
caldwell & coulson ( 1986 ) noted that their cepheids were distributed along a plane inclined with respect to the sky plane ; they are not all at the same distance , but neither are they in a thick distribution liable to generate microlensing .
graff ( in preparation ) , in an analysis of the eros 2 cepheid database , suggest that the extra dispersion in cepheid magnitudes in the smc may be due to a larger intrinsic dispersion for cepheids in the smc compared to the lmc ( which had been used as a standard ) , or perhaps to larger differential reddening .
the cepheid samples may be heavily influenced by cepheids in the wing of the smc .
although the wing does indeed have a large line of sight distance dispersion , it is not substantial enough in the simulation to contribute strongly to the microlensing rate .
our optical depth calculation of @xmath29 is nearly an order of magnitude lower than that calculated ( from only 1 event , macho-97-smc-1 ) by the eros experiment of @xmath36 ( palanque - delabrouille 1998 ) and is also lower than the first order estimate by the macho collaboration of @xmath37 based on both events ( alcock 1998 ) .
since these optical depth calculations are based on only a few events , the discrepancy is at most marginally significant .
the 90% lower limit on the eros optical depth is only @xmath38 ; the corresponding macho value is @xmath39 .
if , however , several more smc events are detected , thereby confirming that the optical depth is larger than @xmath40 , then we will have learned that the gn model contains deficiencies in its representation of the smc .
either the bar is actually thicker than in the gn simulation , or the wing is more substantial . the length of the initial bar used in the simulation was about 8 kpc ; if the bar had maintained its original configuration to the present day , it could be more highly inclined to the sky plane while preserving its projected dimensions , thereby giving a greater thickness along an individual line of sight .
a more extended bar could be realized if the encounter between the magellanic clouds a few @xmath41 yr ago was a less disruptive one . a lower mass for the lmc , a higher mass for the smc , and greater separation between the magellanic clouds at closest approach would all serve to reduce the strength of tidal forces disrupting the edges of the bar .
another possibility to consider is that intense star formation occurred in the spiral arms generated by the tidal interaction , contributing to the optical depth towards the smc centre .
a hydrodynamical simulation which includes star formation would be needed to investigate this possibility . the eros group has suggested that measurements of microlensing in the smc could be interpreted as showing that there is no foreground halo population of lenses ( afonso 1998 ) .
they predicted that future smc events would continue to have long time scales as observed in the first smc event . in that case , since smc events would have different time scales from those observed in the lmc , they could not be due to the same foreground halo population .
our simulation suggests that the situation will not be so clear . as seen in fig.[thatfig ] , the predicted range of time scales from a realistic velocity dispersion and a realistic mass function is broad enough to encompass short time scale events such as those found in the lmc ( marked as crosses ) .
thus , some of the subsequent smc events should have the short time scales seen towards the lmc .
we expect that several smc events will be needed before one can conclusively say that events towards the two galaxies come from different distributions .
this result is not specific to our simulation ; any reasonably broad mass function and transverse velocity distribution should give rise to an equally broad distribution of eintein radius crossing times .
in addition , if our low estimate of the optical depth of the smc is correct , then current experiments will not provide a large enough sample of future events to analyze in this manner .
the gn simulation does not lend support to zhao s ( 1998 ) suggestion that tidal debris from the most recent collision of the smc and the lmc could be responsible for the lmc microlensing events .
nevertheless , kunkel ( 1997 ) have suggested that repeated encounters between the smc and the lmc could form a polar ring around the lmc ; it is feasible that such a structure might be responsible for microlensing .
thus , a simulation with an earlier staring epoch ( the gn simulation only considers the two most recent smc - lmc encounters ) might slowly build up debris bound to the lmc .
exploratory simulations that we have performed with up to four previous encounters do not support this hypothesis ; any nascent ring structure tends to be disrupted by tidal interactions with the milky way .
however , it is conceivable that the gn simulation may not accurately track the past history of the smc - lmc interaction , and that a different orbital configuration may reproduce a large enough cloud of debris along the line of sight to the lmc to generate microlensing .
if we are correct in claiming that the self - lensing optical depth towards the smc is low , then in the near future , present generation experiments will not detect a large number of self - lensing events .
if there is a foreground galactic halo population of lenses , then these lenses will be detected in significant numbers , and will tend to have the short time scales seen in the lmc events .
however , if we are incorrect , i.e. , the optical depth towards the smc is high , then there will be a large background of smc self - lensing events .
a good fraction of these events will have the short time scales seen in the lmc events , and will be indistinguishable from a putative population of foreground halo lenses . ultimately , improved observations ( stubbs 1998 ) may be necessary to resolve the question of the nature of lensing events in the magellanic clouds .
better seeing and larger telescopes will allow the detection of many more events .
better photometry will allow the detection of small changes in the shape of the microlensing light curve which will provide further information on the location of the lenses .
the gn model of the smc is consistent with the hypothesis that the two microlensing events observed towards the smc are due to self - lensing .
the time scales and proper motions in the model are consistent with those observed towards the smc .
the optical depth in the model , @xmath42 , is much lower than the most likely value suggested by the experiments , but still consistent at the 90% confidence level owing to the paucity of events .
this low optical depth is due to the low line of sight thickness of the model .
the gn simulation does not explain the observed microlensing towards the lmc .
nowhere in the simulation outside the smc is the debris dense enough to generate a large optical depth .
tidal debris from the most recent smc - lmc collision is unlikely to account for the lmc optical depth .
owing to the low smc optical depth , suggesting that the number of future events will be low , and the breadth of the time scale distribution , which suggests a high probability that at least some of these future events will have the same time scales as the lmc events , present generation experiments alone are unlikely to resolve the question of whether the smc is lensed by a halo lensing population .
_ acknowledgements : _ the authors would like to thank andrew gould for useful discussions and a careful reading of the manuscript .
ltg acknowledges the sun moon university research centre for financial support of this work .
afonso , c. ( eros ) , 1998 , a&a , 337 , 17 | we analyse an _ n-_body simulation of the small magellanic cloud ( smc ) , that of gardiner & noguchi ( 1996 ) to determine its microlensing statistics .
we find that the optical depth due to self - lensing in the simulation is low , @xmath0 , but still consistent ( at the 90 % level ) with that observed by the eros and macho collaborations .
this low optical depth is due to the relatively small line of sight thickness of the smc produced in the simulation .
the proper motions and time scales of the simulation are consistent with those observed assuming a standard mass function for stars in the smc .
the time scale distribution from the standard mass function generates a significant fraction of short time scale events : future self - lensing events towards the smc may have the same time scales as events observed towards the large magellanic cloud ( lmc ) .
although some debris was stripped from the smc during its collision with the lmc about @xmath1 yr ago , the optical depth of the lmc due to this debris is low , a few @xmath2 , and thus can not explain the measured optical depth towards the lmc . |
quantum entanglement is considered as a key resource in quantum computation , quantum cryptography and quantum information processing @xcite .
one of the most important problems in quantum information theory is whether it is possible to convert an entangled state , that is being shared between two spatially separated parties , to another by applying only local quantum operations on each subsystem .
if two parties are also allowed to communicate classically this entanglement manipulation scheme is called local operations and classical communication ( locc ) @xcite .
two pure states can be obtained with certainty from each other by means of locc if and only if they are related by local unitary operations .
the condition of certainty can be removed to allow probabilistic conversion of states through stochastic local operations and classical communication ( slocc ) @xcite .
this coarse - graining simplifies the equivalence classes labeled by continuous parameters in case of the local unitary operations . with the consideration of slocc
, two states are said to have same kind of entanglement if an invertible local operation ( ilo ) relating them exists @xcite . in case of pure two - qubit states , there is a single equivalence class , i.e. , all entangled states are equivalent to epr state @xmath0 under slocc @xcite . for three qubits , it was shown that entangled states can be converted either to the ghz state @xmath1 , or to the w state @xmath2 by the application of slocc operations @xcite .
in other words , three qubits can be entangled in two inequivalent ways .
although two and three qubit entanglement is relatively well understood , classification problem becomes complicated for multiqubit systems starting from four qubits since continuous parameters are needed to label the equivalence classes @xcite .
the classification of four - qubit entangled states has also been done yet there are different ( but complementary ) points of view on the criteria of classification @xcite .
there are no complete classifications for five or more qubits .
however , for @xmath3-qubit exchange symmetric states , slocc classification has been obtained with the help of majorana representation of symmetric states @xcite . despite the fact that finding an ilo which relates two states with some non - zero probability is sufficient to show the equivalence of these two states under slocc , the success probabilities of transformations have fundamental operational importance in quantum information processes . for pure bipartite states ,
the transformations relating two states of the same class with the greatest probability of success have been found both in the cases of allowing and forbidding classical communication between the parties @xcite .
while the complete solution of the problem is not known for the states involving three or more qubits , there are several works in the literature providing partial solutions @xcite . in this work , flip and exchange symmetric ( fes ) states , which are invariant when two qubits are interchanged or when all 0s ( 1s ) are changed to 1s ( 0s ) , are considered .
it has been recently shown that many - qubit fes states constitute a set of curves in the hilbert space and equivalence classes of these states under ilos can be determined in a systematic way for an arbitrary number of qubits @xcite .
the main purpose of the present work is to investigate the optimal local fes transformations relating two multiqubit fes states assuming that spatially separated parties are only allowed to apply one - shot local operations on their subsystems , i.e. , they are not allowed to make use of classical communication .
although the coordination of local operations by the assistance of classical communication has been shown to enhance the power of transformations in certain cases @xcite , it has also been noted that classical communication is expensive in some situations @xcite .
necessary and sufficient conditions for the entries of a two by two matrix in order for the matrix to be an element of a single qubit quantum operation can be obtained directly from the probability - sum condition of quantum measurements .
consider two operation elements @xmath4 and @xmath5 , which are @xmath6 matrices , and the quantum operation @xmath7 performed on a single qubit .
the only requirement on the operation elements to come from a positive operator - valued measure ( povm ) is the normalization condition that @xmath8 , where @xmath9 denotes the @xmath6 identity matrix and @xmath4 and @xmath5 are defined as @xmath10 for diagonal elements , the normalization condition requires that @xmath11 with @xmath12 therefore , @xmath13 and @xmath14 .
the schwarz inequality @xmath15 implies that @xmath16 s we obtain @xmath17 denotes the determinant of the operation element @xmath4 . on the other hand , if @xmath18 are given with @xmath19 and @xmath20 , eq .
( 7 ) will ensure that @xmath14 .
hence , eq . ( 2 ) together with eq .
( 7 ) give the necessary and sufficient conditions for a @xmath6 matrix to be a valid operation element .
moreover , one can show that the inequalities given by eq .
( 2 ) are guaranteed to be satisfied provided that eq .
( 7 ) holds and @xmath21 .
thus , the constraints can be simplified as latexmath:[\[\begin{aligned } given an operation element @xmath4 with its corresponding probability of success @xmath23 , the entries of @xmath4 can be multiplied by a complex number @xmath24 to increase the success probability of the transformation by a factor of @xmath25 . in this case
, one has latexmath:[\[\begin{aligned } it is obvious from the above expression that the greatest value of @xmath25 , and consequently , of @xmath27 will be obtained when @xmath25 is equal to the right hand side of eq .
( 9 ) . as a result , for the transformations having the maximum probability of success , eq .
( 8) becomes latexmath:[\[\begin{aligned } if eq .
( 10 ) is not satisfied for a given operation element , one can easily scale it to give the optimal probability by multiplying with the maximum allowed value of @xmath25 given by eq .
( 9 ) . a special class of transformation schemes , called one successful branch protocols ( osbp ) , for the distillation of entangled states have been considered in the literature @xcite .
this scenario involves @xmath3 parties performing a unique two outcome povm , whose operation elements are constructed in a way that after each povm , one of the two possible resulting states contains no @xmath3-partite entanglement . for each party
, this restriction mathematically implies that @xmath29=0 $ ] , assuming the successful branch is realized by the application of @xmath4 . the fact that this condition is nothing but the equality part of eq .
( 10 ) guarantees the optimality of osbp in the case of one - shot quantum operations . finally , it is also possible to show that the necessary and sufficient conditions given by eq .
( 8) are equivalent to the constraint on povm elements that the eigenvalues of @xmath30 should be less than or equal to one .
a simple and systematic method for the classification of fes states has been recently obtained @xcite .
it has been noted that imposing flip and exchange symmetry on the system drastically simplifies the form of operators .
thus , fes ilos can be written as @xmath31 where @xmath32 . assuming that @xmath33 is a normalized @xmath3 qubit fes state , all equivalent normalized fes states
can then be obtained as @xmath34 provided that @xmath35 is real .
as @xmath35 changes from @xmath36 to @xmath37 , excluding @xmath38 , @xmath39 traces the curve .
however , if @xmath33 turns out to be an eigenstate of @xmath40 , no fes ilo will alter it or by definition @xmath33 will form an equivalence class by itself .
eigenstates of @xmath41 are of the form @xmath42 where @xmath43 , and number of @xmath44 and @xmath45 states in the kronecker product are @xmath23 and @xmath46 , respectively .
flip symmetric ones are those with even @xmath47 .
eigenvalues are given by @xmath48 and they are @xmath49 fold degenerate .
the eigenstate @xmath50 denotes the fes state obtained by evaluating the symmetric linear combination of degenerate eigenstates corresponding to eigenvalue @xmath51 given by eq .
( 13 ) . in case of three qubits ,
possible even @xmath47 values are 0 and 2 .
while the former corresponds to the separable state @xmath52 , the latter corresponds to the entangled state @xmath53 , which is equivalent to the @xmath54 state .
since @xmath55 can be written as @xmath56 with @xmath57 , it lies on the geodesic connecting the separable @xmath58 state and the entangled fes @xmath54 state . under ilos while @xmath54 ( @xmath59 ) and @xmath58 ( @xmath60 ) are the neighbors of this equivalence class .
( b ) @xmath61 and @xmath62 are the end points of the curves .
the dotted line denotes a portion of the great circle @xmath63 and @xmath64 corresponds to @xmath65 .
the states lying inside the envelope can be generated using @xmath66 with @xmath67 as a representative subset .
( c ) all curves extend between @xmath68 and @xmath69 .
the states lying inside the envelope can be generated using the representative subset @xmath70 with @xmath67 , which is denoted by the dashed line . ]
allowed @xmath47 values for four qubits are 0 , 2 and 4 .
the first and the third are separable @xmath61 and @xmath62 states , respectively .
the only entangled one is @xmath64 which is nothing but @xmath65 ( this state was misprinted as @xmath71 in ref .
@xcite ) in the notation of ref . @xcite where @xmath72 is defined by @xmath73 since there are three distinct eigenvalues , the fes subspace is a sphere .
all curves start and end on @xmath61 and @xmath62 .
expectedly , there exists infinitely many curves corresponding to infinitely many different slocc classes . among the nine classes of four - qubit states , the only fes one is @xmath72 with @xmath74 and @xmath75 , and it represents a great circle on the sphere passing through @xmath64 and making equal angles with @xmath61 and @xmath62 @xcite .
hence , all four - qubit fes states can be generated , by the application of fes ilos , using @xmath63 as a representative subset .
if one specifically wants to deal with the curves lying inside the envelope , then considering @xmath66 with @xmath67 as a representative is sufficient .
when it comes to five qubits , the only separable eigenstate is @xmath69 .
the remaining two are entangled states represented by @xmath76 and @xmath68 .
the fes subspace is again three dimensional and the curves join @xmath69 and @xmath68 . since
all three distinct eigenstates are perpendicular to each other by construction , all curves lying inside the envelope can be generated by the application of fes ilos to the representative subset @xmath70 with @xmath67 .
since fes ilos given by eq . ( 11 ) have a fixed form , one can scale the operators to obtain the optimal local transformations of multiqubit fes states by multiplying the matrices with the greatest allowed value of the scaling factor @xmath77 , which is @xmath78 where @xmath79 . for three qubits ,
if one assumes that the initial state is @xmath55 , then @xmath39 tends to the entangled state @xmath59 as @xmath80 .
however , fig .
2(a ) shows that the maximum probability of obtaining a final state in the close neighborhood of the entangled state @xmath59 decays to zero . on the other hand
, @xmath39 tends to the separable state @xmath60 as @xmath81 .
it can also be seen from fig .
2(a ) that the probability of obtaining a final state in the vicinity of the separable state @xmath60 is at most @xmath82 .
furthermore , the maximum probability of success for transforming an arbitrary initial state @xmath83 with @xmath84 to a final state , which is in the vicinity of the separable state @xmath60 , is examined .
2(b ) displays that the closer the initial state to the entangled state @xmath59 , the more robust it becomes against a possible fes noise source .
class fes states starting from the @xmath55 state .
( b ) maximum probability of obtaining a final state in the vicinity of @xmath60 assuming that the initial state is @xmath83 with @xmath84 .
( c ) maximum probabilities of obtaining four - qubit fes states , under the assumption that the three initial states are @xmath85 ( solid line ) , @xmath86 ( dashed line ) and @xmath87 ( dotted line ) .
( d ) maximum probability of obtaining a final state in the close neighborhood of one of the separable states when the initial state is @xmath66 with @xmath84 .
( e ) maximum probabilities of obtaining five - qubit fes states , under the assumption that the initial states are @xmath88 ( solid line ) , @xmath89 ( dashed line ) and @xmath90 ( dotted line ) .
( f ) maximum probability of obtaining a final state in the vicinity of the separable state @xmath69 when the initial states are arbitrary points on the curves generated from @xmath90 ( dotted line ) , @xmath91 ( dashed line ) and @xmath88 ( solid line ) . ] for four qubits , among the infinitely many curves joining @xmath61 and @xmath62 , three of them are chosen for the investigation of the optimal fes transformations . as @xmath81 ( @xmath80 )
, all three initial states get closer and closer to the separable state @xmath61 ( @xmath62 ) .
2(c ) shows the probability of success of optimal fes transformations , assuming that the initial states are @xmath85 ( solid line ) , @xmath86 ( dashed line ) and @xmath87 ( dotted line ) .
it should be noted that this discussion is fundamentally different from the three - qubit case since the initial states chosen here belong to different fes slocc classes .
2(d ) displays the maximum probability of obtaining a final state in the close neighborhood of one of the separable states when the initial state is @xmath66 with @xmath84 . considering the plots
, one can conclude that the closer the entangled four - qubit fes states to the entangled state @xmath64 , the more robust they become in the sense that the optimum probability of converting them to the states lying in the vicinity of @xmath62 and @xmath62 vanish . for five qubits ,
the number of curves , which corresponds to the number of different fes slocc classes , are also infinite and again only three of them are considered .
while all three initial states tend to the separable state @xmath69 as @xmath81 , they approach to the entangled state @xmath68 as @xmath80 .
2(e ) illustrates the probability of success of optimal fes transformations , assuming that the initial states are @xmath88 ( solid line ) , @xmath89 ( dashed line ) and @xmath90 ( dotted line ) . the asymmetry in the plot is due to the fact that @xmath68 is an entangled state while @xmath69 is a separable state .
2(f ) shows the maximum probability of obtaining a final state in the close neighborhood of the separable state @xmath69 when the initial states are arbitrary points on the three curves generated from @xmath90 ( dotted line ) , @xmath91 ( dashed line ) and @xmath88 ( solid line ) .
consequently , five - qubit entangled states , which are in the vicinity of the curve connecting @xmath76 and @xmath68 , are more robust than other entangled states .
in summary , optimal local fes transformations relating two multiqubit fes states have been investigated by making use of the constraints on the elements of a quantum operation .
it has been shown that some entangled fes states are more robust than others under fes ilos .
namely , for certain initial states , optimal probability of success drops to zero as the final state approaches to a separable state .
for example , among the three - qubit @xmath55-equivalent states , those that are closer to the fes @xmath54 state are more robust than the other members of the class .
although our calculations have been limited to the three , four and five - qubit cases , the generalization of the present work to include @xmath3-qubit fes states is straightforward , since a systematic method has been presented for classifying these states under slocc @xcite .
this work has been partially supported by the scientific and technological research council of turkey ( tubitak ) under grant 107t530 .
the authors would like to thank b. akmak for helpful discussions .
99 r. horodecki , et al . ,
( 2009 ) 865 .
bennett , d.p .
divincezo , j.a .
smolin , w.k .
wootters , phys .
76 ( 1996 ) 722 .
bennett , d.p .
divincezo , j.a .
smolin , w.k .
wootters , phys .
a 54 ( 1996 ) 3824 .
bennett , s. popescu , d. rohrlich , j.a .
smolin , a.v .
thapliyal , phys .
rev . a 63 ( 2000 ) 012307 .
w. dr , g. vidal , j.i .
cirac , phys .
a 62 ( 2000 ) 062314 .
lo and s. popescu , phys .
a 63 ( 2001 ) 022301 .
f. verstraete , j. dehaene , b. de moor , h. verschelde , phys .
a 65 ( 2002 ) 052112 .
l. lamata , j. len , d. salgado , and e. solano , phys .
a 75 ( 2007 ) 022318 .
l. borsten , d. dahanayake , m. j. duff , a. marrani , and w. rubens , phys .
( 2010 ) 100507 .
d. li , x. li , h. huang , and x. li , phys .
rev . a 76 ( 2007 ) 052311 .
t. bastin , s. krins , p. mathonet , m. godefroid , l. lamata , e. solano , phys .
103 ( 2009 ) 070503 .
g. vidal , phys .
83 ( 1999 ) 1046 .
m. a. nielsen , phys .
83 ( 1999 ) 436
. b. he , j. a. bergou , phys .
a 78 ( 2008 ) 062328 .
a. acn , e. jan , w. dr , and g. vidal , phys .
85 ( 2000 ) 4811 .
m liang , c .- z li , phys .
lett . a 308 ( 2003 ) 343 .
a. yildiz , phys .
a 82 ( 2010 ) 012317 .
s. kinta , s. turgut , j. math .
51 ( 2010 ) 092202 .
w. cui , e. chitambar , h .- k .
lo , phys .
a 82 ( 2010 ) 062314 .
w. cui , w. helwig , and h. k. lo , phys .
a 81 ( 2010 ) 012111 .
z. gedik , opt .
commun . 284
( 2011 ) 681 .
lo and s. popescu , phys .
83 ( 1999 ) 1459 . | local quantum operations relating multiqubit flip ( 0 - 1 ) and exchange symmetric ( fes ) states , with the maximum possible probability of success , have been determined by assuming that the states are converted via one - shot fes transformations .
it has been shown that certain entangled states are more robust than others , in the sense that the optimum probability of converting these robust states to the states lying in the close neighborhood of separable ones vanish under local fes operations . |
the penetration of radiation into an optically thick distribution of gas is a feature of many astrophysical systems , ranging from scales as small as those of circumstellar disks to those as large as the damped lyman-@xmath3 absorbers observed along the sightlines to many quasars .
numerical modelling of the propagation of radiation through the gas can greatly aid our efforts to understand the astrophysics of these systems , but frequently proves to be computationally challenging , owing to the high dimensionality of the problem . in the common case in which we have no useful spatial symmetries to exploit and wish to solve for the properties of the radiation field within @xmath4 different frequency bins ,
the computational cost of determining the full spatial and angular distribution of the radiation field is of order @xmath5 , where @xmath0 is the number of resolution elements ( e.g. grid cells in an eulerian simulation , or particles in a smoothed particle hydrodynamics [ sph ] model ) , and where we have assumed that the desired angular resolution is comparable to the spatial resolution . for static problems , where the gas distribution is fixed and we need only to solve for the properties of the radiation field at a single point in time , it is currently possible to solve the full radiative transfer problem numerically even for relatively large values of @xmath0 ( see e.g. * ? ? ?
* who post - process the results of an sph simulation with @xmath6 ) . however , if one is interested in dynamical problems , where the gas distribution is not fixed and the gas and radiation significantly influence one another , then the cost of solving for the radiation field after every single hydrodynamical timestep can easily become prohibitively large ( for a detailed discussion , see @xcite ) .
for this reason , it is useful to look for simpler , more approximate techniques for treating the radiation that have a much lower computational cost , and that can therefore be used within hydrodynamical simulations without rendering these simulations overly expensive .
one common simplification that nevertheless has a reasonably broad range of applicability is to ignore the re - emission of incident radiation within the gas . making this simplification means that rather than solving the full transfer equation , @xmath7 along multiple rays through the gas , where @xmath8 is the specific intensity at a frequency @xmath9 , @xmath10 and @xmath11 are the emissivity and opacity at the same frequency , and @xmath12 is the path length along the ray
, one can instead solve the simpler equation , @xmath13 equation [ full_rt ] has the formal solution @xmath14 where @xmath15 is the specific intensity of the radiation field at the start of the ray ( e.g. at the edge of a gas cloud ) , @xmath16 is the source function , and @xmath17 is the optical depth along the ray .
if @xmath18 and the optical depth is not too large , then it is reasonable to neglect the integral term , in which case we can write @xmath8 as @xmath19 which is the formal solution to equation [ simple_rt ] . by making this approximation
, we therefore reduce the problem to one of determining optical depths along a large number of rays .
often , this problem can then be further reduced to one of determining the column density of some absorber ( e.g. dust ) along each ray .
unfortunately , although these simplifications make the problem easier to handle numerically , they do not go far enough , as the most obvious technique for calculating the column densities integrating along each ray still has a computational cost that scales as @xmath2 and hence is impractical in large simulations .
this motivates one to look for computationally cheaper methods for determining the angular distribution of column densities seen by each resolution element within a large numerical simulation . in this paper
, we introduce a computationally cheap and acceptably accurate method for computing these column densities , suitable for use within simulations of self - gravitating gas that utilize a tree - based solver for calculating gravitational forces .
our method , which we dub , makes use of the large amount of information on the density distribution of the gas that is already stored within the tree structure to accelerate the calculation of the required column density distributions . in the next section ,
we give a description of how our algorithm works , starting with a overview of how tree - based gravity solvers work in section [ overview ] , and then showing how it is possible to implement the method in section [ implement ] .
we then present two stringent tests of the algorithm in section [ tests ] , both of which are typical of the conditions found in contemporary simulations of star formation .
we discuss some of the potential applications of the method in section [ applications ] . in section [ performance ] , we give an overview of the computational efficiency of this scheme , and we summarise this paper in section [ summary ] .
schematic diagram showing how the tree is constructed and used for the gravitational force calculation .
a 3d oct - tree splits each parent node into eight daughter nodes , but in this 2d representation , we show only four of these nodes .
the black lines show the boundaries of the tree nodes that would be constructed for the given ensemble of particles , shown as blue dots .
the regions shaded in red denote the nodes that would be used to calculate the gravitational force as seen by the large blue and orange particle at the bottom of the diagram .
note that in the case where the nodes being used contain only one particle ( a ` leaf ' node ) , the position of the particle itself is used to calculate the gravitational force arising from that node.,width=249 ] schematic diagram illustrating the concept . during the tree
walk to obtain the gravitational forces , the projected column densities of the tree nodes ( the boxes shown on the right ) are mapped onto a spherical grid surrounding the particle for which the forces are being computed ( the `` target '' particle , shown on the left ) .
the tree already stores all of the information necessary to compute the column density of each node , the position of the node in the plane of the sky of the target particle , and the angular extent of the node .
this information is used to compute the column density map at the same time that the tree is being walked to calculate the gravitational forces . provided that the tree is already employed for the gravity calculation , the information required to create the @xmath20 steradian map of the column densities can be obtained for minimal computational cost . ]
tree - based gravity solvers ( e.g. @xcite ) have long been a standard feature of @xmath0-body and smoothed particle hydrodynamics codes ( e.g. @xcite ) .
more recently , their accuracy and speed has also seen them adopted in grid - based codes @xcite . in this paper , we describe a method whereby the information stored in the gravitational tree can be used to construct a @xmath20 steradian map of the column density . by constructing this map at the same time
as the tree is being `` walked '' to determine the gravitational forces , we can minimize the amount of additional communication necessary between cpus holding different portions of the tree . since the structure of the tree , and how it is walked , will be important for our discussion
, we will first give a brief overview of how a tree - based gravity solver works .
for the purpose of this discussion , we consider a solver based on an oct - tree , as used in e.g. the gadget sph code @xcite , although we note that solvers based on other tree structures , such as binary trees , do exist ( e.g. the binary tree employed by @xcite , which later found its way into other high profile studies , such as @xcite and @xcite ) .
a tree - based solver starts by constructing a tree , splitting the computational volume up into a series of nested boxes , or ` nodes ' .
the ` root ' node is the largest in the hierarchy and contains all of the computational points in the simulation
. this large ` parent ' node is then split up into eight smaller ` daughter ' nodes as shown in figure [ fig : tree ] .
the daughter nodes are further refined ( becoming parents themselves ) until each tree node contains only one particle ( illustrated in figure [ fig : tree ] by the blue dots ) .
these smallest nodes at the very bottom of the hierarchy are typically termed ` leaves ' . at each point in the hierarchy , the tree stores the information about the contents of the parent node ( including its position , mass and size ) that will be needed during the gravitational force calculation .
once the construction of the tree is complete , each particle is located in a leaf node situated at the bottom of a nested hierarchy of other nodes .
once the tree is built , it can then be `` walked '' to get the gravitational forces .
the idea behind the speed - up offered by the tree gravity solver over direct summation is very simple : any region of structured mass that is far away can be well approximated as a single , unstructured object , since the distances to each point in the structure are essentially the same . strictly , this is only true if the angular size of the structure is small , and so tree - codes tend to adopt an angle , rather than a distance , for testing whether or not structures can be approximated .
this angle is often referred to as the `` opening angle '' of the tree , and we will denote it hereafter as . to walk the tree to obtain the gravitational force on a given particle ,
the algorithm starts at the root node and opens it up , testing whether the daughter nodes subtend an angle of less than .
if the angle is smaller than , the properties of the daughter nodes ( mass , position , centre of mass ) are used to calculate their contribution to the force . as such
, any substructure within the daughter nodes is ignored , and the mass inside in the nodes is assumed to be uniformly distributed within their boundaries .
if one or more of these nodes subtends an angle larger than , the nodes are opened and the process is repeated on their daughter nodes , and so on , until nodes are found that appear smaller than . to increase the accuracy of the force calculation , the nodes often store multipole moments that account for the fact that the node is not a point mass , but rather a distributed object that subtends some finite angle ( e.g. see @xcite ) .
these moments are calculated during the tree construction , for all levels of the node hierarchy except the leaves , since these are either well approximated as point masses as is the case for a stellar @xmath0-body calculation or are sph particles , which have their own prescription for how they are distributed in space @xcite .
the above method is sketched in figure [ fig : tree ] , which shows the tree structure in black , and the nodes , marked in red , that would be used to evaluate the gravitational force on the large blue particle with the orange highlight . in the cases where the nodes are leaves ( containing only a single particle ) , the position of the particle itself is used . as the total number of force calculations
can be substantially decreased in comparison to the number required when using direct summation , tree - based gravity solvers offer a considerable speed - up at the cost of a small diminution in accuracy .
@xcite showed that for a distribution of @xmath0 self - gravitating particles , the computational cost of a tree - based solver scales as @xmath21 , compared to the @xmath22 scaling associated with direct summation .
they also showed that the multipole moments allowed quite large opening angles , with values as large as 0.5 radians resulting in errors of less than a percent .
our method makes use of the fact that each node in the tree stores the necessary properties for constructing a column density map .
the mass and size of the node can be used to calculate the column density of the node , and its position and apparent angular size allow us to determine the region on the sky that is covered by the node .
note also that column density , just like the total gravitational force , is a simple sum over the contributing material , meaning that it is independent of the order in which the contributions are gathered .
just as the tree allows us to construct a force for each particle , we can also sum up the column density contributions of the nodes to create a @xmath20 steradian map of the column density during the tree - walk .
a schematic diagram of how this works is shown in figure [ fig : treecoldia ] . the target particle the one currently walking the tree , and for which the map is being created is shown as the large dark blue particle on the left . around it
we show the spherical grid onto which the column densities are to be mapped .
we see that the tree nodes , shown on the right , subtend some angle @xmath23 ( which is less than some adopted ) , and cover different pixels on the spherical grid . during the tree
walk , the method simply maps the projection of the nodes onto the pixels for the particle being walked .
schematic showing the overlap between a pixel on the sph particle s _ healpix _ sphere , and the tree node .
the angular size of the pixels and nodes are denoted by @xmath24 and @xmath25 , respectively , and the distances between their centres are given by the orthogonal angles @xmath26 and @xmath27 .
the diagram shows the case when the angle subtended by the tree node is greater than that of the pixels , and the two possible situations that can arise : a ) the pixel and tree node only partially overlap , and b ) the pixel is entirely covered by the tree node . in the former case , we work out the mass in the overlapping area , and convert it to a column density contribution by smearing it over the pixel s area . in the latter case
, the pixel just obtains the full column density of the node . in the case where the angle subtended by the pixels is greater than the tree node ( not shown here ) , then obviously the tree node can also become totally covered by the pixel . in this case , the full mass of the node is smeared out over the pixel s area to define the column density contribution .
full details of how the mapping is done in this implementation are given in section [ implement].,width=307 ] illustration showing how the nodes and pixels are assumed to interact in our implementation of the algorithm .
for each tree node , a new co - ordinate system is created , in which the node s position vector is the @xmath28-axis .
the angular distance between the node centres , can then be described by two orthogonal angles , @xmath26 and @xmath27 , which allows us to define an overlap area @xmath29 .
note that nodes and pixels have an area @xmath30 and @xmath31 respectively .
full details are given in section [ implement].,width=307 ] the details of exactly how the nodes are mapped onto the grid depends on how accurate one needs the column density information to be .
however , it should be noted that the tree structure is only an approximate representation of the underlying gas structure : it distributes the mass in a somewhat larger volume than is actually the case , and as a result , sharp edges tend to be displaced to the boundary of node . as such , column densities from the tree will always be approximate , and so a highly accurate mapping of the node column density projections is computationally wasteful . in
what follows , we will describe a simple implementation of that is both reasonably accurate while at the same time requiring minimal computational cost .
our mapping of the tree nodes to the pixels makes a number of assumptions regarding the shape and projection of the nodes and the pixels .
these are : * the tree nodes are always seen as squares in the sky , regardless of their actual orientation . *
the nodes are assumed to overlap the pixels as shown in figure [ fig : nodepro ] , such that we can define the overlapping region based on simple orthogonal co - ordinates in the plane of the sky .
* we use the _ healpix _ algorithm @xcite to compute pixels that are equidistant on the sphere s surface and that have equal areas .
we show a schematic diagram of the way the nodes are assumed to overlap in figure [ fig : pixdia ] .
the tree nodes are taken to be squares with side length @xmath25 and likewise , the pixels onto which the column densities mapped are assumed to be squares with side length @xmath24 . as shown in the diagram , these dimensions are assumed to be equivalent to the angles subtended by the nodes and the pixels .
overlap requires that @xmath32 and @xmath33 if this is the case , the lengths , @xmath34 and @xmath35 describing the overlapping area are then given by @xmath36 and @xmath37 when the pixels have a smaller angular extent than the nodes ( i.e. @xmath38 ) , or @xmath39 and @xmath40 when the nodes have a smaller angular extent than the pixels ( i.e. @xmath41 ) . by taking the minimum of the expression @xmath42 and either the node or pixel side length , we account for situations such as those shown in the right - hand panel in figure [ fig : pixdia ] , in which either the pixel is totally covered by the node , or the node is totally contained within the pixel .
we can then calculate the contribution of the node to the pixel s column density from @xmath43 this expression is formed by considering the mass in the overlapping area given by @xmath44 .
if the pixel is totally covered by the node , then it gets the full column density of the node .
if the pixel is only partially covered by the node , then the mass in the overlapping region is smeared out over the area of the pixel , to create a new column density . if the node is totally contained within the pixel , then obviously all the mass from the node is smeared out over the pixel s area .
clearly , the ability of @xmath23 and @xmath45 to describe the overlapping area breaks down near the poles , with the extreme case where a pixel directly over either of the poles can not be described by a @xmath35 . to account for this
, we move to a co - ordinate system in which the tree node s position vector , @xmath46 describes a new @xmath28-axis ( @xmath47 ) , such that the node is always located at ( 1 , 0 , 0 ) . to define the other two axis ( the new @xmath48 and @xmath49 axis ) ,
we first define a control vector , @xmath50 , that is close to the node s position vector , displaced by a small amount in @xmath23 and @xmath45 . for the displacement we choose ( somewhat arbitrarily ) that @xmath23 decreases by @xmath51 for @xmath52 and increases by @xmath53 for @xmath54 , and that @xmath45 always increases by @xmath53
we then define the new axis vectors from : @xmath55 @xmath56 @xmath57 @xmath58 the pixel unit vectors @xmath59 are then rotated into this co - ordinate system simply by taking the scalar product with each of the axes : @xmath60 @xmath61 @xmath62 a schematic diagram of how the pixels and nodes are defined in this new coordinate system is given in figure [ fig : nodepro ] . the angular distances @xmath63 and @xmath64 can then be defined simply from : @xmath65 and , @xmath66 to increase the speed of the algorithm , these inverse trigonometric functions can be made into look - up tables . it should be noted that after this rotation , the pixels and even the tree node itself are in general _ not _ aligned as they appear in figure [ fig : nodepro ] . in the above coordinate transform , the control vector @xmath50 determines how the new coordinate vectors @xmath67 and @xmath68 are orientated with respect to the new x axis , @xmath69 ( that is , how @xmath67 and @xmath68 are rotated _ around _
@xmath69 ) .
in fact , it should also be stressed that the _ healpix _ pixels are not aligned as in figure [ fig : nodepro ] _ before _ the rotation , but in fact appear more diamond shaped ( as one can see in the maps in figures 5 - 9 ) .
we found that the exact rotation is typically unimportant for the mapping , provided the angular resolution in the map is not significantly smaller than the opening angle used during the tree - walk .
we discuss this further in section [ tests ] . in our discussion
so far , we have referred only to ` nodes ' , and their properties , but it should be stressed that some of the nodes will be ` leaves ' . in our implementation , the leaves are sph particles , and as was mentioned above , it is customary to use the particle properties directly when evaluating the gravitational forces ( or in our case , the column density ) . as such
, we adopt the exact particle position when considering the leaf nodes . however
, although sph particles have non - uniform radial column density profiles , we do not take this into account in our implementation , but rather treat the sph particles in the same manner as the other nodes , by assuming that they have a uniform column density , and project a square in the sky , rather than a circle . as our node to pixel mapping is based around mass conservation ( the concept behind equation [ sigmaadd ] ) , we can not simply use the smoothing length @xmath70 to define the square ( so @xmath71 in gadget 2 , or @xmath72 for the definition of @xmath70 in most other sph codes ) , but instead have to define @xmath73 to conserve area , giving , @xmath74 our motivation for treating the sph particles in this way , is that working out the true fraction of the sph particle s mass that falls within the pixels is computationally expensive , requiring a numerical integration over the overlapping areas , since the pixel and the sph particle have quite different shapes .
our choice of the _ healpix _ method of pixelating the spheres around the sph particles was motivated by two main factors .
first , the pixels in the _ healpix _ mapping are equal area , which simplifies any comparisons of the pixel properties between different maps .
second , the equal - area property means that increasing the pixel number is equivalent to increasing the angular resolution of the map _ everywhere on the sphere_. this is not the case with the traditional latitude - longitude discretisation , for example , in which the pixels at the poles have a significantly smaller area than their counterparts at the equator .
finally , for our sph code , we use the publicly available code gadget 2 @xcite , which uses an oct - tree .
for this project , where we need to have control over the opening angle @xmath75 to show how it affects the results , we adopt the standard barnes and hut opening criterion @xcite rather than the ` relative error ' criterion suggested by @xcite .
in this section we apply the algorithm to two very different types of test problem . in the first case , we consider a gas cloud that is an isolated sphere , with conditions similar to the dense cores found in the pipe nebula @xcite . for the second test problem , we consider a cloud that is a model of a turbulent molecular cloud , which is representative of the environment in which prestellar cores form .
these two different set - ups are typical of those used in contemporary simulations of star and cluster formation . in what follows
, we will use hammer projections to display the @xmath76 steradian maps of column densities seen from given locations within the cloud .
there are four types of map that we will show .
the first is the ` true ' column density map obtained by summing up the contribution from every single sph particle in the simulation . for this type of map
, it is customary to take into account the radial density profile of the sph particles , as described by their smoothing kernel .
however since this is not done in the implementation we choose not to do this here for the sph maps .
instead we assume that each sph particle has a constant column density defined simply by its mass , and radial extent ( that is , the particle s smoothing length ) .
the second type of map is the ` pixel - averaged ' map , whereby we pixelate the ` true ' map into a set number of _ healpix _ pixels , by averaging over the points from the hammer projection that lie inside each pixel .
this type of map provides a more useful measure than the ` true ' maps , as the results of are also stored on _ healpix _ grids . as such , could be said to be working perfectly if it can recover the same column densities as those shown in the ` pixel - averaged ' maps .
the third type of map is simply the column density map produced by .
finally , our last type of map describes the error in the method .
we define a fractional error @xmath77 for each pixel @xmath78 by @xmath79 where @xmath80 is the column density of pixel @xmath78 in the ` pixel - averaged ' map , and @xmath81 is the column density in pixel @xmath78 recovered by . in our tests , we will also explore the two intrinsic resolutions that are at play in our implementation of .
the first is the number of pixels in the _ healpix _ sphere that surrounds the sph particles , which represents the ability of the sph particle to record the column density information that comes from the tree walk .
the second resolution at play is the opening angle , @xmath75 , as this determines how accurately the tree is forced to look at the structure in the cloud .
together these determine the accuracy and level of detail that is present in the map .
the column density hammer projections for a particle sitting on the edge of a centrally condensed sphere .
the upper left panel shows the column density projection from all sph particles in the simulation volume .
the other panels then show the same @xmath20 steradian map pixelated into 48 , 192 , and 768 _ healpix _ pixels .
the pixel values are simply averages of hammer projection points that lie inside each pixel s boundary.,width=326 ] the maps recovered by our implementation , for the column density distribution shown in figure [ spheretrue ] .
the maps are shown for two different measures of the resolution : the opening angle , @xmath75 , of the tree ( a measure of how well can ` see ' the cloud ) , and the number of pixels in the healpix map ( a measure of how accurately s results are stored).,width=326 ] in the first test problem , we consider a particle located at the edge of a spherical , isothermal cloud with a mean mass density of @xmath82 g@xmath83 , a temperature of 10 k and a mass of 1.33 .
the cloud is modelled with 10,000 sph particles , and hence the mass resolution is comparable to that used in contemporary models of cluster formation ( e.g. @xcite ) .
the cloud is gravitationally unbound , but confined by an external pressure of @xmath84 and has been allowed to settle into hydrostatic equilibrium .
it centrally condenses into a stable bonnor - ebert sphere @xcite with a central density of @xmath85 g@xmath83 and an outer density of @xmath86 g@xmath83 at a radius of 0.09 pc .
the column density map of the sky , as seen by the particle at the edge , is shown in hammer projections in figure [ spheretrue ] . in this figure
, we show the ` true ' column density map , as calculated from the individual sph particles that make up the cloud , and also the map pixelated into 48 , 192 , and 768 _ healpix _ pixels to create the ` pixel - averaged ' maps described above . on the left - hand side of the hammer projections
one can see the high column density of the centrally condensed core of the sphere , and on the right - hand side of each map one can see the edge of the sphere , and the empty void beyond .
although such a simple cloud geometry may seem trivial , it actually represents a stringent test of the algorithm .
first , the tree itself is made up of a series of boxes , and so the intrinsic geometries of the cloud and tree are quite different .
second , we would expect that the rapidly evolving gradient in the column densities associated with the sharp edge of the cloud will be difficult for the tree to capture , as the edges of the nodes will tend to be in a different place , as discussed above . despite these difficulties , the algorithm is able to capture the main features of the cloud fairly well .
figure [ spheretreecol ] shows the representation of the sky maps given in figure [ spheretrue ] for two different tree opening angles , @xmath87 and @xmath88 .
we can see that the column density towards the centre of the cloud is well represented , and that the maps have the same overall features as those in figure [ spheretrue ] : high column density on one side , and a fairly sharp decline on the other side where the column density falls to zero .
although the images in figure [ spheretreecol ] give an idea of the structure and boundaries that is able to reproduce , it is difficult to gauge the quantitative accuracy of the method . a better representation is shown in figure [ sphereerror ] , where we plot the relative error in the maps .
here we see that the error is typically less than 10 percent when the column density is high , but can be as large as around 100 percent when the column density is low , or approaching zero . the high error ( around 50 percent ) in the middle of the map ( and the outer extremities ) comes from the fact that the boundaries of the tree nodes are not necessarily aligned with the edge of the particle distribution . as we increase s ability to see the structure in the cloud , by reducing the opening angle , we see that the error at the boundary decreases .
overall , the best representation of the cloud s boundary ( and indeed the cloud itself ) is found in the 48 pixel map that was run with a tree opening angle of 0.3 .
this is unsurprising , as the low resolution of the pixel - averaged map is also unable to capture the sharp fall in the column density at the cloud s boundary , while at the same time the smaller opening angle ensures that the pixels on the boundary are not assigned mass that belongs to further inside the cloud . in general , we see that the smaller opening angles tend to produce better maps for a given pixellation .
this is expected , since as the opening angle is reduced , the properties of the tree nodes become closer to the actual distribution of the particles .
this is most obviously apparent in the 768 pixel map , where we see that the map obtained for @xmath88 contains artefacts from the underlying boxy structure of the tree , while the @xmath87 map is much smoother . in the maps with a lower number of pixels ,
these features are not so apparent as the structure of tree is more smeared out .
perhaps a more useful measure of the ability of to sample its surroundings is the error in the average column density in the map , as given in table [ meancol ] . for the spherical cloud set - up
, we find that the average column is between 4.2 and 7 percent higher than the average in the true map , with the lowest resolution run ( @xmath88 , n@xmath89 = 48 ) having the largest overall error , and the highest resolution run ( @xmath87 , n@xmath89 = 768 ) having the lowest error .
the fact that these errors are so low reflects the fact that the mean is dominated by the high column density regions , which are recovered well by in all the resolutions we study .
the relative error ( computed according to equation [ frac_error ] ) based on the difference between the maps shown in figure [ spheretreecol ] and the pixellated maps shown in figure [ spheretrue].,width=326 ] the column density sky - map as seen by a low - density particle in a turbulent molecular cloud simulation . as in figure [ spheretrue ] , the upper - left panel is obtained by adding up the contributions from all sph particles in the computational volume ( excluding the particle from which the sky is viewed ) .
the other panels then show a ` pixel - averaged ' view of the cloud , as would be seen if we only had 48 , 192 and 768 pixels in our map.,width=326 ] the left - hand panels show the maps for the turbulent cloud set - up shown in figure [ turbtrue ] , for 48 , 192 and 768 pixels in the map .
all maps were produced using a tree opening angle @xmath87 .
the right - hand panels show the relative error in the maps.,width=326 ] .a summary of the mean column densities in the cloud models presented in sections [ sec : spheretest ] and [ sec : turbtest ] , for both the true map ( the first line ) and each of the maps . for the results we give the number of pixels used in the column density map ( n@xmath89 ) , the opening angle of the tree ( @xmath75 ) , and the percentage error compared to the true map from the sph particles .
note that due to the way the pixel - averaged maps are obtained ( see section [ tests ] ) , their average column density is identical to that in the full sph map , and so we do not include it here . [ cols="^,^,^,^,^ " , ] [ scaling ]
although the parallel efficiency of gadget 2 for this problem is not particularly high to begin with , it is clear that the use of does not have very much influence on the scaling , suggesting that the additional communications overhead is not a significant problem in comparison to the inherent difficulties involved in properly load - balancing a simulation of this type . however , as with any computational method , there are drawbacks to our approach .
one of the main downsides of the method is that it can introduce a significant memory overhead
. the exact memory requirements of can vary considerably , depending on the type of tree employed by the code , how the column density information is being used , and whether the code is parallelized using the message passing interface ( mpi ) protocol , or using the openmp protocol . in gadget 2 for example a code that is mpi parallelized copies of the sph particles on a given cpu are sent to all the other cpus to get the contributions to the gravitational force from the particles that reside there .
an implementation in gadget 2 must then store two copies of the column density map for each particle : one that is broadcast to the other cpus to pick up their contributions , and one that resides on the home cpu that collects the local contributions and stores the final total .
other tree codes parallelized using mpi work differently , sending the necessary information _ from _ the other cpus to the cpu with the target particle , requiring that only one map be stored per particle .
in fact , if the column density map is only needed once for example , to compute a mean extinction then only the particle currently walking the tree needs a column density map . in this case
the information stored in the map can be used at the end of particle s walk , and the map can then be cleared in preparation to be re - used in the next particle s tree - walk . so depending on the application , and on the code , the memory requirements for can be anywhere from one map per parallel task , to two maps per particle .
we have present a new tree - based technique for obtaining a full @xmath76 steradian map of the column densities at every location in numerical fluid simulations .
the method piggy - backs on a tree - based gravitational force calculation , by making use of the information that is already stored in the tree namely the mass , position , and size of the tree nodes to construct a map of the column density distribution in the sky as seen by each fluid element .
as the underlying algorithm is based on the tree , the method inherits the same @xmath90 scaling as the tree code .
the fact that the method makes use of physical quantities that are already stored in the tree means that it is simple to implement , and requires only minimal modification to the underlying tree algorithm . in the case where the tree has been parallelised ,
we find that the inclusion of does not significantly affect the parallel scaling of the code . in this paper
, we describe a simple implementation of that we find to yield column density maps that are accurate to better than 10 percent on average . in this implementation which in our case was made within the publicly available sph code gadget 2 @xcite
we adopt the _ healpix _ @xcite pixelisation scheme to define the pixellated map in which the column densities for each particle are stored . as an example application of
we show how the method can be used to calculate the dust heating of prestellar cores by the interstellar radiation field .
the results are compared with those from the monte carlo radiative transfer code radmc-3d . comparing our lowest resolution results
48 pixels in the @xmath76 steradian _ healpix _ map and a tree opening angle of 0.5 to a 20 million photon packet radmc-3d calculation , we find that the two methods yield radial dust temperature profiles that agree to within 0.5k .
we also discuss some other applications in which we expect to be useful , such as the attenuation of uv radiation and its effect on the chemical and thermal balance of molecular clouds or the x - ray heating of the intergalactic medium .
we would like to thank mordecai - mark mac low , tom abel , gabriel aorve , and lszl szcs , for many interesting discussions regarding the method , and help with assembling the final manuscript .
the authors acknowledge financial support from the landesstiftung baden - wrrtemberg via their program internationale spitzenforschung ii ( grant p - ls - spii/18 ) , from the german bundesministerium fr bildung und forschung via the astronet project star format ( grant 05a09vha ) , from the dfg under grants no .
kl1358/10 and kl1358/11 , and via the sfb 881 `` the milky way galaxy '' , as well as from a frontier grant of heidelberg university sponsored by the german excellence initiative .
the simulations reported on in this paper were primarily performed using the _ kolob _ cluster at the university of heidelberg , which is funded in part by the dfg via emmy - noether grant ba 3706 . | we present _ treecol _ , a new and efficient tree - based scheme to calculate column densities in numerical simulations . knowing the column density in any direction at any location in space is a prerequisite for modeling the propagation of radiation through the computational domain .
_ treecol _ therefore forms the basis for a fast , approximate method for modelling the attenuation of radiation within large numerical simulations .
it constructs a _
healpix _ sphere at any desired location and accumulates the column density by walking the tree and by adding up the contributions from all tree nodes whose line of sight contributes to the pixel under consideration . in particular when combined with widely - used tree - based gravity solvers the new scheme requires little additional computational cost . in a simulation with @xmath0 resolution elements , the computational cost of _ treecol _ scales as @xmath1 , instead of the @xmath2 scaling of most other radiative transfer schemes .
_ treecol _ is naturally adaptable to arbitrary density distributions and is easy to implement and to parallelize , particularly if a tree structure is already in place for calculating the gravitational forces .
we describe our new method and its implementation into the sph code gadget 2 .
we discuss its accuracy and performance characteristics for the examples of a spherical protostellar core and for the turbulent interstellar medium .
we find that the column density estimates provided by _
_ are on average accurate to better than 10 percent . in another application ,
we compute the dust temperatures for solar neighborhood conditions and compare with the result of a full - fledged monte carlo radiation - transfer calculation .
we find that both methods give very similar answers .
we conclude that _ treecol _ provides a fast , easy to use , and sufficiently accurate method of calculating column densities that comes with little additional computational cost when combined with an existing tree - based gravity solver .
methods : numerical radiative transfer |
at present the phase transition signal search for hadrons and nuclei into quark - gluon plasma @xcite and back to hadronization , is in progress .
these studies are carried out at high and low energies @xcite .
the data obtained at working accelerators are checked whether these transitions are possible . the search for the phase transitions
is closely connected with a high ( extreme ) multiplicity region exceeding considerably the mean multiplicity @xcite .
we suggest that the transition to the quark - gluon phase happens in this region , and this phase can be revealed by means of the collective behavior of secondary particles in the proton and nuclear interactions at accelerator u-70 ( ihep , protvino ) .
the search is realized in the extreme multiplicity region because the bose - einstein condensation has been predicted in this very region @xcite .
also the indications were obtained on the formation of ring events ( analogy of cherenkov emission gluons from quarks ) @xcite and on the grouping of secondary particles clusterization @xcite .
now the experimental and theoretical studies are in progress to check the assumption about the increased soft photon ( less than 50 mev ) yield in comparison with the estimations obtained in quantum electrodynamics .
the existing theoretical models and monte - carlo event generators differ very much in their estimations and multiplicity behavior predictions in the high multiplicity region .
we have developed a gluon dominance model ( gdm ) @xcite to describe the extreme multiplicity and mechanism of hadronization .
this model predicts the maximal number of charged and neutral particles at 70 gev which does not exceed 26 and 16 , correspondingly .
multiplicity distributions obtained at mirabel : @xmath0 - @xmath1 gev and @xmath2 - @xmath3 gev , at svd-2 : @xmath4 - @xmath1 gev , the solid curve - gdm prediction.,width=3,height=3 ] the investigations are carried out on the modern experimental setup svd-2 @xcite ( spectrometer with vertex detector ) equipped with a strip silicon detector , a drift tube tracker , a magnetic spectrometer with proportional chambers , cherenkov counter and an electromagnetic calorimeter.we have designed and manufactured a scinillator hodoscope ( high multiplicity trigger ) producing a signal to register events with multiplicity not less than the given level
the so called trigger level . the realization of this project will enable the community to move ahead towards the deeper understanding of the strong interaction nature in the extreme multiplicity region .
since 2005 we have been working with the 50 gev proton beam .
our experimental studies are carried out at svd-2 ( spectrometer with vertex detector ) setup @xcite on u-70 accelerator of ihep ( protvino , russia ) .
this installation consists of the following basic elements : a hydrogen or nuclear target , a precise vertex detector ( pvd ) , a straw tube chamber or a drift tube tracker ( dt ) , a magnetic spectrometer ( ms ) with proportional chambers ( pc ) , cherenkov counter and an electromagnetic calorimeter or gamma - quantum detector ( dega ) .
we manufactured a scintillation hodoscope or high multiplicity trigger which produces a signal to record the events with not lower than the specified multiplicity level .
the main element of svd-2 setup is pvd .
it allows one to reconstruct the interaction vertex with a high degree of accuracy .
it was manufactured on the basis of strip silicon sensors with a step of 25 and 50 @xmath5 m .
it has a set of planes at the following angles : @xmath6 , @xmath7 and @xmath810.5@xmath9 .
we distinguish coordinates @xmath10 , @xmath11 , @xmath12 ( + 10.5@xmath9 ) and @xmath13 ( -10.5@xmath9 ) , respectively .
the oblique planes @xmath14 and @xmath15 necessary for disentangling of tracks in space , were installed in november 2008 . in the previous 2006 and 2007
runs we obtained double multiplicity distributions on each from two projections : _ xoz _ and _ yoz _ , where axis _ z _ is the beam direction . after the 2008 run we performed data processing with two oblique planes and got preliminary single multiplicity distributions using the pvd data ( high multiplicity trigger level was equal to 8) and compared it with mirabel data and gdm . fig .
1 illustrates four multiplicity distributions for comparison : mirabel results at 70 gev ( empty square ) and 50 gev ( full square ) , the last distribution is consistent with svd distribution ( full circle ) and with the gluon dominance model ( the solid curve ) . the suppression of distributions in the small multiplicity region ( n @xmath16 8) is stipulated by the scintillation hodoscope rejections .
this element regulates the high multiplicity event selection .
we have taken into account the efficiency of registration of the high multiplicity events by the pvd .
the monte carlo simulation has demonstrated that the subsequent data processing from dt and ms does not significantly change these distributions .
these detectors can improve , in principle , the registration efficiency for high multiplicity events .
the monte - carlo simulation and comparison with the mirabel data have given the estimations of these losses caused by the limited pvd acceptance as about two charged particles on the average .
these losses have been taken into account in fig . 1 .
now we have designed and debugged new software for track reconstruction by using these additional detectors based on kalman filter technique .
this algorithm can use the tracks reconstructed in pvd and propagate them in dt and ms for a more precise track parameter estimation ( especially momentum ) and then it also finds additional tracks in dt and ms with the hits which were not used in the previous step .
it means that we can find tracks which are invisible at pvd ( because of the restricted acceptance ) and determine the momentum of these particles .
it is also possible to work without pvd information for track reconstruction but only in dt and ms .
the examples of the track reconstructed events at this energy by means of this algorithm at three modules of dt and sixteen pc , are shown in figs . 2 and 3 .
example of mc simulated event and its reconstructed tracks at dt and ms by using a new algorithm.,width=3,height=2 ] example of mc simulated event and its reconstructed tracks at dt and ms by using a new algorithm.,width=3,height=2 ]
the search for the collective phenomena in the high multiplicity region is in progress .
we consider that they can be observed in this region .
the indications on the ring events were obtained in pa interactions at high ( more than 18 ) multiplicity @xcite .
it is interesting to analyze pp interactions and compare them with nuclear interactions .
let us determine @xmath17 as an angle between tracks of the primary and secondary particles . using the pvd data for different multiplicity intervals ( small and high ) which can be considered as a certain value of the impact parameter
, we have discovered a two - hump structure .
it was revealed in the extreme multiplicity region . in fig .
4 we give the @xmath17 distribution for the events with more than 8 charged particles .
the solid line is the result of the polynomial approximation of the seventh order .
we see two different picks .
these picks are absent in the case of small multiplicity - not more than 8 charged particles ( fig . 5 ) . in this picture
we compare these @xmath17 - distributions .
the monte - carlo simulation was carried out and high multiplicity events were selected .
the reconstruction of these events has shown the absence of the two - hump structure for the @xmath17 - distributions ( fig .
6 ) .
if we assume that this two - hump structure is caused by cherenkov gluon radiation , then it is possible to use the formula by dremin @xcite for the index of refraction .
the value of @xmath18 determines angle @xmath17 between the direction one of the maximal humps and the primary track : @xmath19 rad . according to the dremin theory for gluon rings @xcite , @xmath20 where @xmath21 and @xmath22 is the index of refraction . at beam momentum @xmath23=50 gev and proton mass @xmath24=0.938 gev from ( 1 )
we obtain the experimental value of the index of refraction at 50 gev : @xmath25 using the formula by dremin we obtain the following : @xmath26 @xmath27 where @xmath28 - mass of the parton , @xmath29 - the number of scatterers within a single nucleon ( conventional number is equal to 7 ) , @xmath30 is the ratio of real and imaginal parts of the scattering amplitude of partons , @xmath31 - energy of the parton .
the parton mass and its momentum is replaced by the values for proton : @xmath32 and @xmath33 .
after that we get @xmath34 we can reach the agreement with our experimental result @xmath35 if @xmath36 = 4.6 @xmath37 or 0.92 fm for parton scatters .
a possibility of of forming bose - einstein condensation ( bec ) formation in the extreme multiplicity region , has been shown in @xcite by begun and gorenstein .
it is known that pions ( charged and neutral ) are copiously formed at u-70 energies .
they are bosons .
their momenta are approaching to zero at high multiplicity and the bec can form .
the pion number fluctuations will be a prominent signal in the bec - point .
they predict that the scaled variance of neutral and charged pion - number fluctuations @xmath38 in the vicinity of bec - line , have an abrupt and anomalous increase .
our project is aimed at checking this prediction in the experiment .
our collaboration is preparing to check this prediction experimentally . for this purpose
we have selected the high multiplicity events to determine the number of @xmath39 ( photon ) for every of them .
then the variance of the number of particle fluctuations of the both neutral and charged pions may give a signal about the bec formation or its absence .
we are sure that the extreme multiplicity studies are very interesting and useful .
the recovering of the two - hump structure , bec formation , search for the turbulence and different collective phenomena will be carried out at lhc and other centers .
we thank all participants of svd collaboration for active and fruitful work on the project .
the authors express deep recognition to i.m .
dremin for stimulating discussion before these studies .
these investigations have been partially supported by rfbr grants @xmath40 and @xmath41 .
99 andronic a braun - munzinger p and stachel j ( 2006 ) nucl phys a**772**:167 - 199 .
cleymans j and redlich k ( 1999 ) phys rev c * 60*:054908 .
becattini f , manninen j and gazdzicki m ( 2006 ) phys rev c * 73*:044905 .
oeschler h ritter h and nu xu ( 2009 ) arxiv:0908.1771[nucl - ex ] .
avdeichikov _ et al .
_ , proposal `` termalization '' ( 2005 ) jinr - p1 - 2004 - 190 .
. begun and m.i .
gorenstein ( 2007 ) phys lett b * 653*:190 - 195 .
kokoulina e and kutov a ( 2008 ) phys atom nucl 71:1543 - 1551 .
kokoulina e kutov a and nikitin v ( 2007 ) braz j phys * 37*:785 - 787 .
kokoulina e and nikitin v ( 2005 ) in _ proceedings of 17th ishepp .
jinr , dubna _ 319 - 326 , 327 - 336 ; the 7th international school - seminar on actual problems of high - energy physics , gomel , belarus , jul 28- aug 8 2003 , arxiv:0308139[hep - ph ] ; kokoulina e aip conf proc ( 2006 ) * 828*:81 - 86 .
aleev a et al svd-2 collaboration nonlinear dynamics and applications ( 2006 ) * 13*:83 - 93 .
kokoulina e , kutov a and riadovikov v ( 2009 ) phys of atom nucl * 72*:184 - 188 .
dremin i ( 1981 ) phys lett b * 102*:40 - 42 ; dremin i ( 2007 ) rom rep phys * 59*:977 - 985 ; dremin i ( 2008 ) acta phys polon supp * 1*:641 - 646 ; dremin i m ( 2009 ) arxiv:0903.2941v2[hep - ph ] . | extreme multiplicity studies at 50 gev in pp interactions are discussed .
preliminary multiplicity distributions at u-70 ( ihep , protvino ) energy have been obtained for more than 20 charged particles .
a new elaborated algorithm for the track reconstruction in a drift tube tracker and magnetic spectrometer , has been checked .
the collective behavior of secondary particles is manifested in these interactions in the extreme multiplicity region .
for the first time the ring events in pp interactions have been observed in this region . a possibility of detecting the bose - einstein condensation detection is discussed . |
superfluid dynamics of interfaces between bose - einstein condensates ( becs ) of dilute atomic gases has gained active interest , both in experimental @xcite , and theoretical works @xcite .
these studies started with the experimental realization of a multicomponent bec @xcite and the theoretical investigation of the planar interface structure @xcite . at present , a good deal of the research focuses on the hydrodynamic instabilities at the interfaces in becs , such as the kelvin - helmholtz instability , the rayleigh - taylor ( rt ) instability , the ferrofluid normal - field instability , and the richtmeyer - meshkov ( rm ) instability @xcite .
a common feature of these fundamental instabilities in the classical gas dynamics is the production of vortices at the unstable interface . in that sense ,
dynamics of an interface between two quantum fluids has some unique features with quantized vorticity being , probably , the most specific and the most interesting among them . at the same time
, there is also much similarity between the classical and quantum gas dynamics , which allows `` borrowing '' some classical results and applying them directly or with modifications to the quantum case , see refs . @xcite .
such `` borrowing '' is especially typical in studies of the linear instability stages , for which the difference between the classical and quantum cases is expected to be minimal .
still , the problem of rigorous derivation of such effects in bec systems from the basic principles of quantum theory remains open for most cases .
the derivation is needed not only to justify the extrapolations from the classical to quantum gas dynamics , but also to find limits of such extrapolation as well as to indicate intrinsic quantum effects involved into the problem . as an example , we point out the study of the interface waves in a system of two becs undertaken in ref . @xcite . apart from common capillary waves anticipated from the classical gas - dynamics and related to in - phase perturbations of two becs , ref .
@xcite demonstrated also the possibility of quantum counter - phase perturbations , for which the local interpenetration depth of two wave functions for different becs oscillates in time .
thus , one purpose of the present paper is to provide rigorous derivation of the dispersion relations for some fundamental quasi - hydrodynamic instabilities in a two - component bec driven by a time - dependent force . as a particular experimental realization of such a force , sasaki et al .
@xcite suggested a system of two phase - segregated , interacting becs with different spins placed in an external magnetic field gradient .
this is possible , for example , for two magnetic sublevels of the @xmath0 hyperfine state of @xmath1 with @xmath2 and @xmath3 .
starting from the gross - pitaevskii ( gp ) action for the two component bec system , we derive the equations of motion of small perturbations of the interface .
the other purpose of the present paper is to demonstrate the diversity of hydrodynamic effects in a system of two separated becs under the action of a time - dependent force .
recently , within classical gas dynamics , there has been increased interest in the problems of the modified rt and rm instabilities subject to time - dependent gravity , e.g. see @xcite .
the configuration of time - dependent gravity was discussed in the context of specially designed laboratory experiments @xcite , in the scope of the problem of inertial confined fusion @xcite and in relation to flame dynamics @xcite . in particular , stabilization of the rt instability by an oscillating high - frequency addition to the gravity acceleration
has been proposed both in the traditional rt configuration @xcite and for the inertial confined fusion @xcite .
development of the parametric instability at a flame front under the action of acoustic waves and possible stabilization of the hydrodynamic flame instability has been investigated experimentally and theoretically in refs .
@xcite . here
we show that similar effects take place in becs . apart from the rt and
rm instabilities considered before in @xcite , we demonstrate also the possibility of dynamical stabilization of the rt instability by high - frequency oscillations , as well as triggering of the parametric instability
. another interesting option , which has not been considered yet even in the classical gas - dynamics corresponds to a gaussian stochastic force , which may be markovian or non - markovian with zero or non - zero time correlations , respectively . in the present paper
we show that the markovian stochastic force does not produce any average effect on dynamics of the interface , while the non - markovian force leads to exponential growth of perturbations .
the paper is organized as follows . in section
ii we formulate the model . in section iii
we consider a planar ( one - dimensional ) quantum system with a magnetic gradient pushing the becs towards each other , for the case of both constant and time - dependent external forces . in section
iv we derive the equation , which describes the dynamics of small interface perturbations . in section
v we consider interface dynamics under the action of different forces , including a constant force , a harmonic force , a pulse force , and a noisy gaussian force .
the results of the paper are summarized in section vi .
the system of two becs at zero temperature is described by two macroscopic wave functions @xmath4 , and the mean - field gp lagrangian @xcite : @xmath5 , \label{eq1}\end{aligned}\ ] ] where @xmath6 are external potential energies of the bec components .
the tildes in eq .
( [ eq1 ] ) designate dimensional variables , which will be scaled in the following . keeping in mind a system with two magnetic sublevels of @xmath7 , we assume the same particle masses for both components @xmath8 and the same interaction parameters @xmath9 .
the interaction parameters @xmath10 , @xmath11 , @xmath12 are related to the respective s - wave scattering lengths @xmath13 by @xmath14 . a dimensionless interaction parameter is introduced through the definition @xmath15 the two components are separated in space if @xmath16 , which implies @xmath17 , see @xcite . throughout this paper
we use the condition of weak separation @xmath18 .
modification of all results of the present paper to the case of different becs is straightforward .
the unperturbed flat interface between two components of bec is located in the @xmath19 plane with @xmath20 indicating the plane of symmetry , fig .
the components @xmath21 and @xmath22 are characterized by their projection of the atom magnetic moment on the @xmath23-axis @xmath24 , respectively . using the states
@xmath25 and @xmath26 of the atoms makes it possible to apply the stern - gerlach force , pushing the components in the opposite directions perpendicularly to the magnetic moment quantization axis .
we use a magnetic field whose magnitude has the gradient @xmath27 , which is uniform in space but may be time - dependent . for the case of @xmath28 atoms the stern - gerlach potential energy in eq.([eq1 ] ) reads @xmath29 where @xmath30 is the bohr magneton , and @xmath31 is the land@xmath32 factor . the chemical potentials @xmath33 in eq.([eq1 ] ) define normalization of stationary wave functions @xmath34 , and are equal to energy per particle of @xmath35-th condensate , which is made up of the interaction energy due to s - wave scattering and of energy of the particle in the external potential . in order to study low - energetic hydrodynamic processes in becs
we assume that energy per particle is changed negligible due to the external potential , and for normalization one can use value of the chemical potential for @xmath36 in an infinite system .
such system is characterized by a uniform density sufficiently far from the interface ( _ i.e. _ on length scales much larger than the interface thickness ) ; then the chemical potential is given by a usual result for a uniform bec with number density @xmath37 , @xmath38 .
however , real becs are always finite ; if the length scale of our system along @xmath39-axis is @xmath40 ( each condensate is of size @xmath41 ) , then the condition that the external potential is a small perturbation , reads @xmath42 dimensionless units are introduced by scaling coordinates by @xmath43 , time by @xmath44 , and atom number density ( or concentration ) by @xmath45 .
the dimensionless magnetic force acting on the components is @xmath46 the dimensionless gross - pitaevskii lagrangian for two macroscopic wave functions @xmath47 at zero temperature reads @xmath48 , \label{eq3}\end{aligned}\ ] ] and equations of motion are obtained from the action principle @xmath49,\ ] ] where the action is @xmath50 .
the result is the coupled gp equations : @xmath51{\psi _ 1 } , \\
i{\partial { \psi _ 2 } \over { \partial t } } = \left [ { - { \delta \over 2 } - 1 + z b(t ) + { { \left| { { \psi _ 2 } } \right|}^2 } + \left ( { 1 + \gamma } \right){{\left| { { \psi _ 1 } } \right|}^2 } } \right]{\psi _ 2}.\end{aligned}\ ] ]
the present work is devoted to multidimensional instabilities bending an interface between two bec components in an external magnetic field gradient . as the first step in the study
, we have to specify the planar ( one - dimensional , 1d ) quasistationary state , which is influenced by a spatially uniform external field . in the subsections
a and b we consider the cases of constant and time - dependent external fields , respectively . here
we study the 1d hydrostatic state of becs , when the constant external force @xmath52 is balanced by internal forces of the condensates .
let us assume that such a state was created by adiabatic switching on of the external field , and no perturbations bend the interface .
the steady state is described by two real - valued functions @xmath53 , and the system of dimensionless gp equations reads @xmath54 the time - independent force pushes the bec components towards each other when @xmath55 . with our choice of equal masses and scattering lengths for both components ,
the problem possesses the symmetry @xmath56 . in the case of zero magnetic gradient
the density profiles have been found in @xcite within the limit of weak separation @xmath57 as @xmath58^ { - 1}}.\ ] ] we calculate the wavefunction of condensate 1 in the bulk of condensate 2 . according to eq .
( [ eqarray7 ] ) , the wave functions penetrate into each other with the characteristic depth @xmath59 , see also @xcite , which means that @xmath60 for @xmath61 .
the density profile of the bulk is described by the thomas - fermi approximation ( tfa ) , which means that @xmath62 for @xmath61 . with these approximations eq.([eqarray7 ] ) becomes : @xmath63{\psi _ { 01}}.\ ] ] the type of solution to eq.([eq8 ] ) changes from exponentially decaying to oscillating in space , as @xmath39 becomes larger than @xmath64 \approx \gamma/(2 b_0)$ ] .
however , the magnitude of the spatial oscillations depends on the value @xmath65 : it decreases exponentially with increasing @xmath66 .
therefore , the spatial oscillations become significant when @xmath66 becomes comparable with @xmath67 , or when @xmath68 ; in this case the tfa becomes invalid , and eqs.([eqarray7]),([eqarray8 ] ) must be solved exactly .
the spatial oscillations may be interpreted as quantum interference of matter waves , caused by interplay between the non - linearity and the dispersion ( the kinetic term ) .
this phenomenon is beyond the scope of the present work , and it will be studied in detail elsewhere .
thus , here we use parameters satisfying the condition @xmath69 or the magnetic gradient @xmath70 satisfying @xmath71 which means that the spatial oscillations of the hydrostatic profiles may be neglected .
it follows from eq .
( [ eq9 ] ) that the critical magnetic gradient @xmath72 depends on the difference between inter- and intra - species interaction parameters as @xmath73 , and on density via the chemical potential since @xmath74 in the tfa for an infinite system in a sufficiently weak external field .
therefore , the effect of quantum interference becomes stronger at lower densities of the becs .
equation ( [ eq10 ] ) yields the dimensionless condition of a well - defined interface between two condensates subject to the particular external force . in the system of two magnetic sublevels of @xmath75
we have @xmath76 ( see @xcite ) , and the typical density is @xmath77 . the numerical solution to eqs .
( [ eqarray7 ] ) , ( [ eqarray8 ] ) for these parameters shows that the spatial oscillations are noticeable at @xmath78 ; so we take @xmath79 .
numerical solutions for the wave functions near the interface are presented in fig . 2 for @xmath80 , plots ( a ) and
( b ) , respectively . the dashed lines in fig .
1(a ) show the approximate analytical solution of eq.([eq7 ] ) obtained for the case of zero magnetic field .
as we can see , the analytical solution eq.([eq7 ] ) fits the numerical results quite well even in the case of non - zero magnetic field . ;
@xmath81 , figures ( a ) and ( b ) , respectively .
the dashed lines show the analytical solution eq .
( [ eq9 ] ) .
the inset in figure ( b ) presents the airy function of the second kind.,width=345,height=451 ] since we are interested in the development of hydrodynamic instabilities for time - dependent external forces , we need to know how a harmonic force influences the planar density profiles of becs .
similar to the constant force studied above , for time - dependent forces we also consider rather small magnetic fields , so that the respective 1d bogoliubov excitations may be interpreted as linear acoustic waves .
another type of a time - dependent force studied in this paper is a pulse force corresponding to a weak shock in hydrodynamics .
still , it is known from hydrodynamics that a weak shock may be also described as a linear acoustic wave @xcite .
for this reason , in the present subsection we study modifications of the bec density in the 1d geometry due to the harmonic force . because of the problem symmetry with respect to @xmath20 , we consider a one - component semi - infinite bec for @xmath82 confined by a wall at @xmath20 .
the magnitude and frequency of the force determines the energy of the generated bogoliubov excitations . in this subsection
we express density and velocity perturbations , @xmath83 and @xmath84 , in terms of magnitude and frequency of the external force , and find the restrictions on the driving force , for which the density perturbations are small @xmath85 . in the case of a harmonic force ,
the dimensionless external potential is @xmath86 , where @xmath87 is the force amplitude and @xmath88 is the frequency .
our model system is then described by a time - dependent gp equation @xmath89 where @xmath90 for @xmath82 and @xmath91 for @xmath92 . since the trap potential @xmath93 experiences discontinuity at the `` wall '' , then
the wave function and density are specified only for @xmath92 , no healing of the condensate wave function is required near the wall in tfa , and equilibrium is specified by the uniform density @xmath94 . in that case
the velocity perturbation @xmath95 is zero at the wall @xmath96 the perturbations of density and velocity , @xmath83 and @xmath84 , may be expressed in terms of the perturbation wave function @xmath97 @xmath98 where @xmath99 , and @xmath100 , @xmath101 are time - independent functions .
since the perturbations are small , we find @xmath102 @xmath103 where @xmath104 denotes complex conjugate .
then the boundary condition at the wall reads @xmath105 we linearize the gp equation ( [ eq11 ] ) , @xmath106 and obtain the bogoliubov equations @xmath107 @xmath108 which , in turn , may be reduced to @xmath109{\partial \over { \partial z}}\left ( { u - v } \right ) - { \rm{\omega } } { b_s}.\ ] ] the solution to eq .
( [ eq15 ] ) is a superposition of a running wave @xmath110 and spatially uniform oscillations @xmath111 , @xmath112 where @xmath113 are the respective amplitudes .
the running wave is the bogoliubov elementary excitation ( the quantum acoustic wave ) with the wave number @xmath114 determined via the bogoliubov spectrum @xmath115 it follows from the boundary condition at the wall that @xmath116 and @xmath117.\ ] ] the perturbations of density follow from the 1d continuity equation @xmath118 as @xmath119 then , the condition of weak perturbations @xmath85 becomes @xmath120 , or in the dimensional form @xmath121 where @xmath122 and @xmath123 are the dimensional wave number and frequency of the induced excitations , coupled by the bogoliubov spectrum ( [ eq16a ] ) .
in the low energy limit @xmath124 , the boguliubov excitations are sound waves with velocity @xmath125 ( unity in our dimensionless units ) .
then the restriction on the amplitude of the magnetic force reads @xmath126 the criterion ( [ eq19 ] ) may be interpreted as the condition of incompressible condensates with the speed of sound much larger than any other parameter of velocity dimension involved in the problem . a more general condition ( [ eq18a ] ) takes into account the bogoliubov spectrum for higher energies .
we solved the bogoliubov equations for a model infinite system .
the case of a finite system , confined by two walls , is solved analogously , and the boundary condition ( [ eq13 ] ) should be imposed for both walls .
real systems are always finite , and therefore the maximal wavelength is limited by the size of the system @xmath41 . in the low - frequency limit condition ( [ eq18a ] ) is equivalent to the energetic condition ( [ eq2b ] ) . in this limit ,
the bogoliubov excitations are sound waves , with maximal wavelength of order @xmath41 ; substituting @xmath127 to ( [ eq2b ] ) , and using the sound spectrum , one obtains ( [ eq19 ] ) .
we now consider the dynamics of the interface between two infinite becs in the presence of an external force , which may be time - dependent , @xmath128 . similarly to refs .
@xcite we use a variational approach .
we represent the wave functions in terms of density @xmath129 and velocity potential @xmath130 as @xmath131 , so that the action for the system of two becs is @xmath132 + z\,b(t ) \left ( n_2 - n_1 \right ) + \nonumber\\ \left . {
1 \over 2}\left(n_1 ^ 2 + n_2 ^ 2 \right ) + \left ( 1 + \gamma \right)n_1 n_2 \right\ } .
\label{eq20}\end{aligned}\ ] ] for the variational calculation , we employ an ansatz that is able to describe bending and compression of the interface between the two condensates .
we are interested in perturbation wavelengths much larger than the interface thickness , @xmath133 , so that the density deformations may be treated as a bending of the unperturbed profile as a whole , @xmath134,\ ] ] where @xmath135 are the stationary density profiles for the becs .
the variational functions are taken to be of the form of fourier modes with separation of space and time dependence , @xmath136 @xmath137 in classical hydrodynamics , perturbations of both components of an inert interface are related by the continuity condition @xmath138 . in the quantum gas - dynamics of two becs
this is not necessarily the case , since we may have interpenetration of the wave functions .
the lagrange equations with respect to the variables @xmath139 are @xmath140 . \label{eq23}\ ] ] these lagrange equations are satisfied if the respective continuity equations are @xmath141.\ ] ] first , the time - dependent terms on each side have to be equated , which , up to constant factors , leads to @xmath142 this reflects the fact that @xmath139 and @xmath143 are collective canonically conjugate variables . for the functions
@xmath144 , ref.@xcite gave the approximate solution @xmath145 the same solution holds in the presence of a small external force under the conditions specified in sec .
we take eq .
( [ eq26 ] ) as our ansatz for @xmath144 .
more elaborate expressions for @xmath146 , and @xmath147 may provide a smooth transition in the interface region ; however , the ansatz eq .
( [ eq26 ] ) yields reliable analytical results , as we will demonstrate below by comparison to the numerical solution to the problem .
we now derive the lagrange equation for @xmath148 .
no time derivatives in @xmath143 appear in the lagrangian , so the lagrange equation reads @xmath149 where we introduce @xmath150 to label the nonvanishing terms of @xmath151 ; the numbering @xmath152 in @xmath150 refer to the respective terms in eq .
( [ eq20 ] ) . for convenience , we assume a finite size of the system , @xmath41 , @xmath153 , though the final result does not depend on the size restrictions .
we have @xmath154 the next nonvanishing term comes from the zero - point motion energy , @xmath155= \nonumber\\ & & { 1 \over 2}{\delta \over { \delta { \zeta _ j } } } \int { d^2}\textit{\textbf{r}}\,{\left ( { { \partial \over { \partial z}}\sqrt { { n_j } } } \right)^2}\left [ { 1 + { { \left ( { { \partial \over { \partial x}}\delta { z_j } } \right)}^2 } } \right ] .
\label{eq29}\end{aligned}\ ] ] the first term within the parentheses yields a constant independent of @xmath156 , and the second term equals @xmath157 the next non - vanishing term corresponds to the perturbation of the potential energy of the system , @xmath158 it is more convenient to calculate these terms for @xmath159 and for @xmath160 separately . using the coordinate transformation @xmath161 we modify the last expression as @xmath162,\end{aligned}\ ] ] @xmath163.\end{aligned}\ ] ] the integrals can be calculated as @xmath164 @xmath165 the next term of @xmath166 does not contribute to the equations of motion , while the last ( seventh ) term @xmath167 is important .
analysis of the terms calculated so far shows that the variables @xmath168 are governed by two coupled oscillator equations , and the coupling happens because of @xmath167
. since the equations for @xmath169 and @xmath170 are identical , the normal modes may be in - phase , @xmath171 , and counter - phase , @xmath172 .
first , let us calculate @xmath167 for the case of @xmath171 , which we denote by @xmath173 .
in this case @xmath174 , and we find @xmath175 next we consider the case of counter - phase oscillations with @xmath176 and @xmath177 , which leads to @xmath178 to calculate @xmath179 in the last expression , it is easiest to assume @xmath180 small and taylor expand @xmath181 to second order , which results in @xmath182 the last result may be transformed to a symmetric form similar to @xcite : @xmath183 taking into account all the terms obtained above we write equations of motion for @xmath168 .
the integrals are calculated using the density profile ( [ eq7 ] ) @xmath184 @xmath185 @xmath186 we designate the interface perturbations of the in - phase oscillations by @xmath187 and obtain the final equation @xmath188 = 0.\ ] ] in the opposite case of the counter - phase perturbations @xmath189 we find @xmath190k = 0.\ ] ] in the limiting case of @xmath191 our theory reproduces the results of @xcite . within the limit of a weak field , @xmath192 , and long wavelength perturbations @xmath193 , the terms in the parenthesis of eq .
( [ eq41 ] ) are of different orders of magnitude , with the term @xmath194 dominating . for this reason , to the leading terms , the counter - phase mode is not affected by the external potential , and we do not consider it in the present work . in the following , we work only with the in - phase mode omitting the subscript `` in '' as @xmath195 , and use eq .
( [ eq40 ] ) as the starting point for the subsequent analysis .
in this section we demonstrate the diversity of hydrodynamic effects in a system of two separated becs under the action of a time - dependent force . in particular , we are interested in the rt instability developing due to a constant force , in the rm instability for a pulse force , in the parametric instability for an oscillating force , and in the interface dynamics under the action of a stochastic gaussian force . here
we consider a constant force @xmath196 , which drives the rt instability and capillary - gravitational waves in the unstable or stable configurations , respectively .
constant coefficients of eq .
( [ eq40 ] ) lead to harmonic oscillations of the interface @xmath197 with the dispersion relation @xmath198 which may be written in dimensional form as @xmath199 where @xmath200 and @xmath201 are the dimensional perturbation frequency and wave number .
note that eq .
( [ eq42 ] ) has the same form as the dispersion relation for capillary - gravitational waves in classical hydrodynamics @xcite . when the external magnetic field is zero
, we find capillary waves with frequency related to the wave number as @xmath202 in the case of positive gradient of the magnetic field the dispersion relation eq .
( [ eq42 ] ) yields @xmath203 for perturbations with sufficiently long wavelengths , which corresponds to the rt instability @xmath204 with the growth rate depending on the wavelength as @xmath205 and with the cut - off wave number @xmath206 according to eq .
( [ eq42 ] ) , the maximal growth rate corresponds to @xmath207 .
figure 3 compares the analytical results of eq .
( [ eq42 ] ) , ( [ eq43 ] ) to the numerical solution to the linearized gp equations .
the numerical solution is obtained by diagonalization of the linearized gp equations , approximated by finite difference method on a grid of 3335 points on the interval @xmath208 $ ] in the dimensionless coordinates .
figure 3 shows good agreement of the analytical theory with the numerical results . the instability growth rate eq .
( [ eq45 ] ) agrees also with the formulas employed in @xcite . ) , and presented by lines .
dashed lines show real parts of the bogoliubov eigenvalues ( waves ) , and solid lines correspond to the imaginary parts ( instability ) .
the markers show respective numerical solution to the linearized gp equations ; filled markers correspond to the instability regime and empty markers correspond to waves.,width=336,height=240 ] though the present work considers only the linear stage of the rt instability , for illustrative purposes , it is worth presenting the results obtained for the nonlinear stage .
for example , fig .
4 shows the development of the rt instability at the nonlinear stage in a confined system of two becs obtained numerically using the methods of @xcite . in the numerical solution ,
we begin with a system of @xmath209 atoms of @xmath210 , equally split between the two states , with @xmath211 and @xmath212 .
we use the `` pancake '' trap geometry with @xmath213 and @xmath214 .
we solve the coupled gp equations numerically to find the ground state of the system .
the two components occupy the regions with @xmath215 and @xmath82 , respectively
. a perturbation of size @xmath216 is used , and the field of view for all images is @xmath217 square .
the magnetic field of magnitude @xmath218 is added to the system , directed so that the two components are pushed toward each other .
the system is shown at times 15.9 , 19.08 , 22.26 , 25.44 ( ms ) after the magnetic gradient is added .
the nonlinear stage may be described qualitatively as development of the initial sinusoidal perturbations into a mushroom structure with subsequent generation of quantum vortices
. more detailed investigation of the nonlinear stage of the rt instability will be presented elsewhere .
similar results may be found also in @xcite . in this subsection
we consider an oscillating force @xmath219 , where @xmath52 and @xmath87 are some constant amplitudes and @xmath88 is the frequency of the force .
we show that an oscillating force leads to two interesting effects : 1 ) high - frequency oscillations may stabilize the rt instability produced by the constant background force @xmath52 ; 2 ) the oscillating force may trigger a parametric instability even in the absence of any constant acceleration , e.g. for @xmath220 .
we start with the first effect , which is similar to the mechanical phenomenon known as the kapitsa pendulum @xcite .
we assume that the frequency @xmath88 of the external force is much larger than the rt instability growth rate @xmath221 , that is , @xmath222
. then we separate the slow averaged terms ( labeled `` a '' ) and fast oscillating terms ( labeled `` f '' ) in the solution to eq .
( [ eq40 ] ) as @xmath223 , where @xmath224 and @xmath225 change slowly on the time scale of high - frequency oscillations @xmath226 . following the calculation method of @xcite ,
we take into account the relation @xmath227 and modify eq .
( [ eq40 ] ) to @xmath228\ ] ] @xmath229.\ ] ] separating the terms oscillating with different frequencies we find the relation between the amplitudes of the slow and fast terms @xmath230 and reduce eq .
( [ eq40 ] ) to @xmath231 equation ( [ eq49 ] ) describes stabilization of the rt instability by high - frequency oscillations for @xmath232 thus , stabilization of the rt perturbations of wave number @xmath233 by the high - frequency term is expected for the oscillation amplitude @xmath234 which may be written in dimensional units as @xmath235 , for @xmath236 , @xmath237 and @xmath238.,width=336,height=192 ] the stabilization of the rt instability by high - frequency oscillations is illustrated in figure 5 , which presents the results of direct numerical simulation of eq .
( [ eq40 ] ) for the amplitude ratios @xmath239 , for the stationary component of the magnetic field @xmath236 , the wave number corresponding to the maximal growth rate @xmath237 and frequency of the oscillating magnetic field component @xmath238 . in agreement with the theoretical predictions , unstable exponential growth of perturbations ( in average ) for @xmath240
is replaced by waves at the interface for sufficiently large amplitudes of the oscillating part of the magnetic field , @xmath241 .
figure 6 shows stability limits in the parameter space of the scaled wave number @xmath242 and the oscillation amplitude @xmath243 ; the shading indicates the absolute value of the instability growth rate in the unstable region .
figure 7 shows the rt instability growth rate in the presence of the external high - frequency magnetic field plotted versus the perturbation wave number for different values of the external field .
and the oscillation amplitude @xmath243 , where @xmath236 , @xmath244 .
shading indicates value of the instability growth rate in the unstable region , eq .
( [ eq49]).,width=278,height=230 ] ) , in presence of external high - frequency field with @xmath244 vs perturbation wave number for different values of @xmath245 and @xmath236.,width=336,height=230 ] the second effect related to the oscillating external field is the parametric instability , which is known from mechanics @xcite and which has been encountered in gas dynamics , e.g. , of combustion systems @xcite .
the constant part of the `` acceleration '' is not required to obtain the parametric instability and it may be omitted in the calculations by taking @xmath220 .
similar to @xcite we look for a solution to eq .
( [ eq40 ] ) in the form @xmath246\exp \left ( { { \rm{\alpha } } t } \right)$ ] .
mark that the growth of the parametric instability is accompanied by oscillations of the perturbation with frequency @xmath247 , i.e. half of the driving force frequency .
substituting @xmath248 into eq .
( [ eq40 ] ) we obtain @xmath249 \\ + { \alpha _ 0}{\rm{\omega } } \left [ { - { \xi _ 1}\sin \left ( { { { { \rm{\omega } } t } \over 2 } } \right ) + { \xi _ 2}\cos \left ( { { { { \rm{\omega } } t } \over 2 } } \right ) } \right ] \\ = { b_s}k\cos \left ( { { { { \rm{\omega } } t } \over 2 } } \right)\left [ { { \xi _ 1}\cos \left ( { { { { \rm{\omega } } t } \over 2 } } \right ) + { \xi _ 2}\sin \left ( { { { { \rm{\omega } } t } \over 2 } } \right ) } \right].\end{aligned}\ ] ] separating the terms with @xmath250 and @xmath251 and omitting higher - frequency terms , we reduce the equation to the linear algebraic system for the amplitudes @xmath252 and @xmath253 @xmath254 solving the system we obtain the dispersion relation @xmath255 substituting @xmath256 into eq .
( [ eq52 ] ) we find the limiting oscillation amplitude @xmath87 required to induce the parametric instability @xmath257 in dimensional variables the last relation reads as @xmath258 where the dimensional frequency of capillary waves is @xmath259 and amplitude @xmath87 of the external high - frequency force with @xmath260 .
shading indicates value of the instability growth rate in the unstable region , obtained from eq .
( [ eq52]).,width=278,height=230 ] ) , as function of @xmath233 for different amplitudes of the external oscillating magnetic force @xmath261 with @xmath260.,width=336,height=192 ] minimal ( infinitely small ) critical amplitude of the magnetic field corresponds to the perturbations with the wave number determined by the equation @xmath262 .
therefore , when applying an external magnetic force @xmath263 to the system of two becs , we should expect resonance pumping of the capillary waves with frequency @xmath247 and the wave number @xmath264 .
if capillary waves of another non - resonant wave number dominate initially , then pumping of the waves requires finite ( not necessarily small ) amplitude of the driving force .
the instability domain is shown in fig .
8 in the parameter space of the force amplitude and the wave number with the shading indicating the perturbation growth rate .
figure 9 presents the dispersion relation for the parametric instability ( the growth rate versus the wave number ) for different amplitude values of the oscillating magnetic field . within the context of quantum gases ,
the rm instability produced by a magnetic pulse @xmath265 has been discussed recently in ref .
@xcite , which indicated considerable differences in the instability development in becs in comparison with the traditional configuration of classical gases separated by an inert interface . here
we analyze the linear interface dynamics by a more general ( matrix ) method of solving eq .
( [ eq42 ] ) applicable to a wide class of time - dependent forces .
particularly , the matrix method will be extended to the case of stochastic forces in the next subsection .
we introduce the vector function @xmath266 and rewrite eq .
( [ eq42 ] ) as @xmath267 where the matrix may be split into a constant and time - dependent parts .
first , eq .
( [ eq55 ] ) is solved for the case @xmath268 , which yields @xmath269{\zeta}_{v}(0 ) \nonumber \\ \nonumber
\\ = \left ( { \matrix{{\cos \left ( { { { \rm{\omega } } _ c}t } \right ) } & { \omega_c^{-1}\sin \left ( { { { \rm{\omega } } _ c}t } \right ) } \cr { - { { \rm{\omega } } _ c}\sin \left ( { { { \rm{\omega } } _ c}t } \right ) } & { \cos\left({{{\rm{\omega}}_c}t}\right ) } \cr } } \right){\zeta}_{v}(0).\end{aligned}\ ] ] for convenience , we introduce the designations @xmath270 @xmath271 we look for a general solution to eq .
( [ eq55 ] ) in the form @xmath272 and reduce eq .
( [ eq55 ] ) to @xmath273 where @xmath274 @xmath275 the solution to eq.([eq57 ] ) is the time - ordered exponential @xmath276 , e.g. see @xcite , @xmath277 \equiv 1 + \nonumber \\
\sum \limits_{n = 1}^ { + \infty } \int \limits_0^t d{t_1 } \int \limits_0^{{t_1 } } d{t_2 } \cdots \int \limits_0^{{t_{n - 1 } } } d{t_n}{\bf{f}}({t_1}){\bf{f}}({t_2 } ) \cdots { \bf{f}}({t_n}).\end{aligned}\ ] ] we solve eq .
( [ eq57 ] ) , go back to the original variable and obtain the general solution to eq .
( [ eq55 ] ) @xmath278 on the basis of the general solution , we may recover the results of ref .
@xcite on the rm instability triggered by a pulse @xmath279 . in this case
all terms of time - ordered exponential commute with each other and we obtain @xmath280 @xmath281 which yields the final solution @xmath282}^2 } } \sin \left ( { { { \rm{\omega } } _ c}t + { \varphi _ d } } \right)\ ] ] with the new oscillation phase determined by @xmath283 thus , the general solution eq .
( [ eq59 ] ) reproduces the analytical results of ref .
@xcite given by eq .
( [ eq60 ] ) .
particularly , eq . ( [ eq60 ] ) describes the jump in the amplitude and phase of a standing capillary wave because of the pulse for a non - zero bending of the interface , @xmath284 . in the opposite case of @xmath285 and
@xmath286 the pulse does not influence the interface dynamics .
( [ eq60 ] ) shows that the transfer of magnetic energy of the pulse to superfluid oscillations is the most efficient when the deviation of the interface just before the pulse is maximal .
the linear analysis works rigorously as long as the perturbation amplitude after the pulse remains sufficiently small .
once it becomes large enough , nonlinear effects dominate , and the interface evolves into a mushroom pattern with detachment of droplets as discussed in @xcite .
figure 10 illustrates the development of the rm instability in becs at the nonlinear stage obtained using the numerical methods similar to @xcite .
we use the same system of becs as in fig .
a magnetic pulse of duration @xmath287 and size @xmath288 is applied to the initial capillary wave seen in fig .
the following graphs correspond to the time instants 2.39 ; 4.77 ; and 7.16 ms after the pulse ( figures b , c , d , respectively ) .
development of the rm instability at the initial stages ( e.g. fig .
10 b ) resembles the rt instability shown in fig .
however , the difference between two instabilities is noticeable at later stages shown in fig .
10 ( c , d ) . in particular
, we observe the detachment of droplets in the rm instability in agreement with the model presented in ref .
@xcite .
finally , we consider the interface dynamics under the action of a gaussian random force @xmath289 with mean zero @xmath290 , where the brackets imply averaging over a statistical ensemble of realizations .
the force is switched on at @xmath291 .
because of the gaussian property , the force is characterized only by variance @xmath292 and by the correlation time @xmath293 .
the case of zero @xmath293 is known as the markovian case , and in this case averaging may be performed exactly . on the contrary ,
the case of non - zero @xmath293 is not solvable in general , but an analytical approximate solution may be obtained in the limit of small correlation time @xmath294 . in order to obtain an averaged form of eq .
( [ eq40 ] ) we employ the general solution ( [ eq59 ] ) . since eq .
( [ eq59 ] ) is a time - ordered exponential with non - commutative kernel at different moments of time , we use the cumulant method for averaging @xcite . within this method ,
if @xmath295 is a stochastic gaussian operator with mean zero @xmath296 , then the time - ordered exponential of @xmath295 is averaged by the formula @xmath297}\right\rangle=
\exp \left [ { \int \limits_0^t d\tau \int \limits_0^\tau ds \left\langle{\mathbf{g}(\tau ) \mathbf{g}(s)}\right\rangle}\right].\ ] ] in the markovian case this formula is exact , but in the non - markovian case this is an approximation valid when @xmath293 is small compared to the intrinsic time scale of the system , @xmath294 , and the noise force is sufficiently weak to be treated as a perturbation @xcite .
the kernel of the right - hand side of eq.([eq62 ] ) is the cumulant series terminated at the first non - vanishing term ( the first - order term vanishes because @xmath296 ) .
we note that all cumulants in eq .
( [ eq62 ] ) vanish identically for @xmath298 because of the gaussian property , but they are non - zero for finite @xmath293 due to non - commutativity of the kernel in eq .
( [ eq59 ] ) .
we start with the markovian case when the autocorrelation function is given by the delta - function @xmath299 using eqs .
( [ eq59 ] ) , ( [ eq62 ] ) , we find that the averaged solution to eq .
( [ eq40 ] ) reads @xmath300 _ i.e. _ the markovian stochastic force does not influence averaged dynamics of the interface .
now we consider a non - markovian exponentially correlated noise with autocorrelation function given by @xmath301 which reduces to the markovian case in the limit of zero correlation time @xmath302 .
we average eq .
( [ eq40 ] ) using the novikov theorem @xcite : @xmath303 we calculate the functional derivative in eq.([eq66 ] ) using the general solution eq.([eq59 ] ) @xmath304 ( { t , s } ) \mathbf{\zeta}_{v } ( 0 ) \nonumber\\ & & = \theta(t - s)\theta(s){{\bf{g}}_0}(t - s)k\textbf{v}_{0}{\bf{g}}_0^ { - 1 } ( { t - s } ) \zeta_{v } ( t ) \nonumber\\ & & + { \rm{\delta } } \left [ { { b_r } } \right]\left ( { t , s } \right)\zeta_{v } ( 0 ) , \label{eq67}\end{aligned}\ ] ] where @xmath305 is the heaviside step function , and @xmath306\left ( { t , s } \right)$ ] is a functional arising from the non - commutativity of the kernel of the exponential in eq.([eq67 ] ) .
the functional @xmath306\left ( { t , s } \right)$ ] may be dropped in the present limit of weak noise .
we obtain from eq .
( [ eq67 ] ) @xmath307_{11 } } \left\langle { \zeta ( t ) } \right\rangle + \right .
\nonumber\\ & & \left.\left[\textbf{v}(t - s)\right]_{12}{d \over { dt}}\left\langle { \zeta ( t ) } \right\rangle\right\}. \label{eq68}\end{aligned}\ ] ] substituting the operators @xmath308 and @xmath309 to eq .
( [ eq68 ] ) , and using eq .
( [ eq66 ] ) we find @xmath310 calculating the integrals on time scales much larger than the correlation time of the noise @xmath311 , we arrive to an equation describing the averaged interface dynamics @xmath312 where @xmath313 is the average growth rate and @xmath314 modifies the frequency due to noise .
the solution to eq .
( [ eq70 ] ) grows exponentially in time as @xmath315.\ ] ] thus , a stochastic non - markovian force leads to unstable growth of the average perturbation amplitude with modified frequency of the capillary oscillations . ,
@xmath316 and @xmath317 .
( b ) the respective numerical solution to eq .
( [ eq40 ] ) .
the dashed line shows amplitude of the averaged analytical solution , eq .
( [ eq73]).,width=336,height=192 ] and @xmath318 and @xmath319 .
( b ) the respective numerical solution to eq .
( [ eq40 ] ) ( two realizations ) .
the dashed line shows amplitude of the averaged analytical solution , eq .
( [ eq73]).,width=336,height=192 ] the analytical theory for interface dynamics under the action of a stochastic force is compared in figs .
11 , 12 to the numerical solution to eq .
( [ eq40 ] ) calculated for @xmath320 .
we note that real stochastic processes always have finite frequency band width , though it may be very large .
therefore , any real stochastic process is characterized by some non - zero correlation time leading to exponential growth of amplitude of the oscillator eq.([eq40 ] ) , which means that the markovian case may be recovered only as a limit of @xmath321 .
figure 11 shows the numerical solution to eq .
( [ eq40 ] ) for the stochastic force with @xmath322 and @xmath323 , which may be considered as a good approximation for the markovian case .
the dashed line represents the envelope function of the averaged amplitude of the interface oscillations predicted by the theory , eqs .
( [ eq71 ] ) , ( [ eq73 ] ) . in the markovian case
, the averaged amplitude does not change , which is supported by the numerical results shown in fig .
figure 12 corresponds to a non - markovian case with @xmath324 , @xmath318 and @xmath319 .
the numerical solution to eq .
( [ eq40 ] ) is shown by the solid lines for two realizations of the noise , while the dashed line represents envelope of the averaged amplitude increasing exponentially in time according to the analytical theory eqs .
( [ eq71 ] ) , ( [ eq73 ] ) . in agreement with the theoretical predictions ,
the numerical solution also demonstrates exponential increase of the perturbation amplitude .
we have studied the dynamics of an interface in a two - component bec driven by a spatially uniform time - dependent force .
experimentally , such a force may be produced by a magnetic field gradient acting on the system of two interacting becs with different spins @xcite . by applying the variational principle to the gp lagrangian
, we have derived the dispersion relation ( or the respective ordinary differential equations ) for linear waves and instabilities at the interface .
we have considered different time - dependent forces leading to a diverse collection of dynamical effects : ( i ) a constant force pushing the components towards each other generates the rt instability for a certain domain of perturbation wave numbers . because of the rt instability , small perturbations grow exponentially in time without oscillations .
( ii ) a sinusoidal force with frequency @xmath88 leads to the parametric instability at the interface .
the critical strength of the force required to drive the instability depends on the perturbation wave number .
this strength is infinitely small for the intrinsic capillary waves at the interface with frequency equal to @xmath247 . because of the parametric instability , small perturbations not only grow exponentially in time , but also oscillate with frequency @xmath247 .
( iii ) a pulse force leads to the quantum counterpart of the rm instability . within the linear approximation
, the force produces a discontinuous jump in the amplitude and phase of the capillary waves at the interface .
( iv ) a non - markovian stochastic external field ( with non - zero correlation time ) drives the instability at the interface accompanied by oscillations .
the growth rate of the instability is determined by the variance of the driving force and by the correlation time .
the oscillation frequency is shifted in comparison to the intrinsic frequency of the capillary waves . a markovian force on average
does not lead to any effect .
this research was supported partly by the swedish research council ( vr ) and by the kempe foundation .
calculations have been conducted using the resources of high performance computing center north ( hpc2n ) . c. j. myatt , e. a. burt , r. w. ghrist , e. a. cornell , and c. e. wieman , phys .
lett . * 78 * , 586 ( 1997 ) .
e. g. m. van kempen , s. j. j. m. f. kokkelmans , d. j. heinzen , and b. j. verhaar , phys .
* 88 * , 093201 ( 2002 ) .
s. b. papp , j. m. pino , and c. e. wieman , phys .
101 , 040402 ( 2008 ) .
s. r. leslie , j. guzman , m. vengalattore , j. d. sau , m. l. cohen , and d. m. stamper - kurn , phys .
a * 79 * , 043631 ( 2009 ) .
r. blaaugeers , v. eltsov , g. eska , a. finne , r. haley , m. krusius , j. ruohio , l. skrbek , g. volovik , phys .
* 89 * , 155301 ( 2002 ) . t. ho and v. b. shenoy , phys .
lett . * 77 * , 3276 ( 1996 ) .
p. ao and s. t. chui , phys .
a * 58 * , 4836 ( 1998 ) .
i. e. mazets , phys .
a * 65 * , 033618 ( 2002 ) .
r. a. barankov , phys .
a * 66 * , 013612 ( 2002 ) .
b. v. schaeybroeck , phys .
a 78 , 023624 ( 2008 ) u. al khawaja and h. stoof , nature * 411 * , 918 - 920 ( 2001 ) .
k. sasaki , n. suzuki , d. akamatsu , and h. saito , phys .
a * 80 * , 063611 ( 2009 ) .
s. gautam and d. angom , phys .
a * 81 * , 053616 ( 2010 ) .
g. volovik , jetp letters * 75 * , 418 ( 2002 ) .
a.bezett , v.bychkov , e.lundh , d.kobyakov , and m.marklund , phys .
a * 82 * 043608 ( 2010 ) .
h. saito , y. kawaguchi , m. ueda , phys .
lett . * 102 * , 230403 ( 2009 ) .
h. takeuchi , n. suzuki , k. kasamatsu , h. saito , m. tsubota , phys .
b * 81 * , 094517 ( 2010 ) g. dimonte and m. shneider , phys .
fluids * 19 * , 304 ( 2000 ) .
r. betti and j. sanz , phys .
97 * , 205002 ( 2006 ) .
g. dimonte , p. ramaprabhu , m. andrews , phys .
e * 76 * , 046313 ( 2007 ) .
k. mikaelian , phys .
e * 79 * , 065303(r ) ( 2009 ) .
g. dimonte and m. shneider , phys .
e * 54 * , 3740 ( 1996 ) .
r. betti , r. mccrory , c. verdon , phys .
71 * , 3131 ( 1993 ) .
s. kawata , t. sato , t. teramoto , e. bandah , y. masubishi , i. takahashi , laser part .
beams * 11 * , 757 ( 1993 ) .
v. bychkov , a. petchenko , v. akkerman , l .- e .
eriksson , phys .
e * 72 * , 046307 ( 2005 ) .
v. akkerman , v. bychkov , a. petchenko , l .- e .
eriksson , combust .
flame * 145 * , 206 ( 2006 ) .
a. petchenko , v. bychkov , v. akkerman , and l .- e .
eriksson , phys .
lett . * 97 * , 164501 ( 2006 ) .
a. petchenko , v. bychkov , v. akkerman , and l .- e .
eriksson , combust .
flame * 149 * , 418 ( 2007 ) .
g. wolf , phys .
* 24 * , 444 ( 1970 ) .
v. bychkov , phys .
fluids * 11 * , 3168 ( 1999 ) .
g. searby , combust .
81 * , 221 ( 1992 ) .
g. searby and d. rochwerger , j. fluid mech .
* 231 * , 529 ( 1991 ) c. j. pethick and h. smith , _ bose - einstein condensation in dilute gases _
( cambridge university press , cambridge , 2008 ) . l. d. landau and e. m. lifshitz , _ fluid mechanics _ , ( butterworth - heinemann , oxford , 1995 ) .
u. al khawaja , c. j. pethick , and h. smith , phys .
a 60 , 1507 ( 1999 ) .
landau and e.m .
lifshitz , _ mechanics _ , ( elsevier , amsterdam , 2007 ) .
r.f.fox , phys . rep . * 48 * , 179 - 283 ( 1978 ) .
a. novikov , zh .
. fiz . * 47 * , 1915 ( 1964 ) [ sov .
jetp * 20 * , 1290 ( 1965 ) ] . | the dynamics of an interface in a two - component bose - einstein condensate driven by a spatially uniform time - dependent force is studied . starting from the gross - pitaevskii lagrangian ,
the dispersion relation for linear waves and instabilities at the interface is derived by means of a variational approach .
a number of diverse dynamical effects for different types of the driving force is demonstrated , which includes the rayleigh - taylor instability for a constant force , the richtmyer - meshkov instability for a pulse force , dynamic stabilization of the rayleigh - taylor instability and onset of the parametric instability for an oscillating force .
gaussian markovian and non - markovian stochastic forces are also considered .
it is found that the markovian stochastic force does not produce any average effect on the dynamics of the interface , while the non - markovian force leads to exponential perturbation growth . |
the electromagnetic fields of ultrarelativistic heavy ions may be treated as fields of virtual photons .
when two nuclei pass by at small impact parameters , the photon field of one nucleus can produce a photonuclear interaction in the other nucleus .
most studies of photonuclear reactions in heavy ion collisions have considered hard interactions , such as heavy quark production@xcite or photonuclear breakup@xcite .
we consider here the exclusive reaction @xmath7 , where @xmath8 is a heavy nucleus , and @xmath9 is a vector meson .
this reaction can proceed via photon - pomeron or photon - meson interactions , where the photon comes from the electromagnetic field of one nucleus , and the pomeron or meson couples to the other nucleus . at sufficiently high energies ,
the momentum transfers from the nuclei are small enough that both couplings are coherent to the entire nucleus .
then , the nucleus remains intact , or nearly intact .
we calculate the production rates and rapidity distributions for the @xmath10 , @xmath11 , @xmath6 and j/@xmath12 .
these reactions will be studied experimentally at the upcoming relativistic heavy ion collider ( rhic)@xcite now under construction at brookhaven national laboratory@xcite , and the large hadron collider ( lhc)@xcite being built at cern . for later reference , table [ rhiclum ] gives the expected luminosities for various beam species at these two colliders .
the calculations are simplest in the rest frame of one of the nuclei , labelled the target nucleus .
the total production is found by taking each nucleus as the target in turn .
this calculation divides into two parts : determination of the photonuclear interaction cross section , and calculation of the effective photon flux from the emitting nucleus . in this frame
, the photon energy @xmath13 is large , and the momentum transfer squared from the target nucleus is small .
the momentum transfer @xmath14 has longitudinal @xmath15 and transverse components @xmath16 , with @xmath17 .
since the initial state longitudinal momentum is small , @xmath16 is the final state momentum transverse to the photon direction .
when @xmath18 , @xmath19 .
the coherent coupling of the photons and pomerons limits their virtuality to @xmath20 , where @xmath21 is the nuclear radius , or about 30 mev / c for heavy nuclei like gold or lead .
this virtuality is small enough that it can be neglected here , with the photon effectively real in the target frame .
the photon flux from a relativistic heavy nucleus is given by the weizscker - williams approach . here
@xmath22 , where @xmath23 and @xmath24 are the energy and mass of the nucleus .
the photon flux at a distance @xmath25 from the nucleus is @xmath26 where @xmath13 is the photon energy , @xmath27 is the nuclear charge , @xmath28 a modified bessel function with @xmath29 .
the target frame @xmath30 corresponds to @xmath31 , where @xmath32 is the lorentz boost in the center of mass frame ( laboratory frame for collider geometry ) . here , and throughout the paper
, we take @xmath33 . in an exclusive interaction
, there can be no accompanying hadronic interactions . in a hard sphere model , this occurs if the impact parameter @xmath34 , where @xmath35 fm .
we use a more accurate approach , calculating , the probability of one or more hadronic interactions as a function of impact parameter .
the nuclear density for a nucleus @xmath8 a distance @xmath36 from its center is modelled with a woods- saxon distribution for symmetric nuclei @xmath37}\ ] ] where the radius @xmath38 and skin depth @xmath39 are based on fits to electron scattering data@xcite and @xmath10 is fixed by the normalization .
this electron scattering data probes the proton distribution in the nucleus .
if the neutron are differently distributed from the protons , this could affect the cross section .. the interaction probability at a given impact parameter is related to the overlap function @xmath40 where @xmath41 and @xmath42 are 2-dimensional vectors perpendicular to the direction of travel @xmath43 .
the nuclear thickness function is @xmath44 the number of nucleon - nucleon collisions follows a poisson distribution with mean @xmath45 , with @xmath46 the total nucleon - nucleon interaction cross section .
the probability of having no hadronic interactions is then @xmath47 we use @xmath48 mb at a beam energy of 100 gev , rising to 88 mb at 2.9 tev@xcite .
we assume that the range of the pomeron or meson field is much smaller than the size of the nucleus @xcite . for @xmath49
, a form factor can be required to accurately model the electromagnetic fields inside the emitting nucleus ; we neglect this tiny correction .
the total photon flux is given by eq .
( [ drphotonflux ] ) integrated over all @xmath25 , modulated by the noninteraction probability .
@xmath50 the @xmath51 integral runs over impact parameter and @xmath25 and @xmath6 over the surface of the target nucleus , for a given @xmath51 . this process averages the field strength over the nucleus , neglecting the variation with @xmath25 .
since the vector meson production is coherent over the entire nucleus , this averaging is appropriate .
the integral is evaluated numerically ; the result is shown in fig .
[ fphotonflux ] .
this flux is approximately equal to the photon flux in the region @xmath52 , which can be found analytically : @xmath53 where @xmath54 .
this approximation is shown by the dotted line in fig .
[ fphotonflux ] .
the @xmath55 cross sections are found with a glauber calculation , using the observed @xmath56 cross sections as input .
data on @xmath56 covers a wide energy range , from low energy fixed target photon beams@xcite to @xmath57 collisions at hera@xcite .
these cross sections may be parameterized as @xmath58 where @xmath59 is the center of mass energy , and table [ herasigma ] gives the constants @xmath60 , @xmath61 , @xmath62 , @xmath63 and @xmath64 , which are determined from fits to data@xcite .
the @xmath61 , @xmath63 term represents the pomeron exchange portion of the cross section .
it rises slowly with energy ( @xmath65 ) .
this slow rise has been attributed to the increasing size and density of the proton ; the increasing @xmath59 probes smaller proton momentum fractions @xmath66 .
the @xmath62 , @xmath64 term , is for meson exchange , primarily the @xmath67 @xcite .
the meson exchange term falls rapidly as @xmath59 rises . for the @xmath6 and @xmath2 ,
meson exchange is heavily suppressed , and the reaction occurs only through pomeron exchange . for the @xmath2 , @xmath68 ;
the rapid rise has been ascribed to either a breakdown of the soft pomeron model or to threshold effects@xcite .
because of this behavior , the @xmath2 calculations must be treated with caution , especially at center of mass energies beyond that probed by hera . to relate this to nuclei
, we make use of the optical theorem and an eikonalization technique .
following vector dominance@xcite , @xmath69 with a similar relationship true for nuclei
. here , t is the squared 4-momentum transfer between the proton and vector meson , @xmath70 is the electromagnetic coupling constant , @xmath71 , and @xmath72 is the vector meson - photon coupling , @xmath73 with @xmath74 the vector meson mass and @xmath75 the leptonic decay partial width .
however , measurements of @xmath72 from leptonic decay widths disagree with photoproduction data@xcite . a correction is required to account for non - diagonal coupling through higher mass vector mesons , as implemented in the generalized vector dominance model ( gvdm ) . for the @xmath4 , gvdm increases
@xmath76 by 20%@xcite . because @xmath76 is used twice , the error largely cancels here .
values for @xmath77 are given in table [ herasigma ] .
using the optical theorem , the total cross section is @xmath78 at @xmath79 gev , we find 24 mb , 26 mb , 12 mb , and 1.0 mb for the @xmath4 , @xmath11 , @xmath6 , and @xmath2 total cross sections respectively .
these values are consistent with theoretical and experimental expectations@xcite .
the scattering cross section from heavy nuclei can be found by a glauber calculation @xmath80 the optical theorem for nucleus @xmath8 and gvdm are then used to find @xmath81 these @xmath82 are shown in fig .
[ dsdtgamma ] for gold nuclei . also shown
are the results of scaling the @xmath3 cross sections by @xmath83 and @xmath84@xcite .
the @xmath4 , @xmath11 , and @xmath6 results vary only slowly with energy . the minimum around photon - proton center of mass energy @xmath85 gev for the @xmath4 and @xmath11 marks the transition from meson dominated to pomeron dominated reactions .
@xmath6 production , which is mediated only by pomerons , shows a monotonic energy dependence . for the light mesons ,
@xmath82 is almost independent of energy , flatter than the corresponding proton functions , which appears unchanged in the @xmath83 and @xmath84 scaling curves . in contrast
, @xmath82 for the @xmath2 rises rapidly with energy .
part of the difference is the different @xmath3 cross section energy dependence , and part is the smaller @xmath2 interaction cross section . because @xmath86 is large , @xmath4 interact near the nuclear surface , and @xmath87 is the geometric cross section @xmath88 which is almost independent of @xmath13 .
however , @xmath89 is much smaller , and the @xmath2 interactions occur throughout the nucleus . as long as @xmath90 , the rising @xmath89 will reflect itself in @xmath91 .
these cases correspond to the black disk and weak absorption limits respectively .
the hera results which provide the high energy @xmath3 input data extend up to about @xmath92 = 180 gev ( @xmath93 tev in the target frame ) , with higher energy points extrapolations .
although the hera data covers the entire rhic energy range , the extrapolations are required for lhc .
the uncertainty is especially problematic for the @xmath2 , where @xmath82 must eventually flatten out , to avoid dominating the total photonuclear cross section .
the total cross section depends on the slope of @xmath94 . for proton targets
, the cross section can be parameterized as @xmath95@xcite .
the slope @xmath51 , and derivative @xmath96 represent the sum of the proton size and the range of the interaction . here , for simplicity we take @xmath97 with little loss of accuracy .
nuclei are much larger than protons , so @xmath51 is dominated by the nuclear size , with @xmath98 and @xmath94 is dominated by the nuclear form factor .
this is important because experimental study of these exclusive reactions depends on their small @xmath16@xcite which is determined by the form factor .
the form factor is the fourier transform of the nuclear density profile .
unfortunately , the woods - saxon distribution does not have an analytic form factor .
instead , for simplicity we approximate the woods - saxon distribution as a hard sphere , with radius @xmath21 , convoluted with a yukawa potential with range @xmath99 fm@xcite .
the fourier transform of this convolution is the product of the two individual transforms : @xmath100 \
\ \bigg[{1\over1+a^2q^2}\bigg ] \label{eqff}\ ] ] figure [ fformfactor ] compares eq .
( [ eqff ] ) with the numerical transform of the woods - saxon distribution for gold .
the agreement is excellent .
in addition to coherent production , there is also incoherent vector meson production , where the photon interacts with a single nucleon in the target .
incoherent interactions have a larger average @xmath14 , and occur at a somewhat lower rate .
they will not be further considered here , but they will be an essential component of any experimental analysis .
the photonuclear cross section is @xmath101 for narrow resonances , @xmath102 . because of its width , the @xmath4 is more complicated , and the cross section must be calculated using a breit - wigner resonance : @xmath103 here , @xmath104 is the pole position of the resonance , @xmath105 the total cross section ( neglecting phase space corrections ) and @xmath106 the final state invariant mass .
the breit - wigner width @xmath107 includes the phase space correction , with @xmath108 @xcite where @xmath109 is the decay pion momentum in the @xmath4 rest frame , with @xmath110= 358 mev / c the pion momentum for @xmath111 and @xmath112 is the pole @xmath4 width .
the observed resonance shape is the convolution of this breit - wigner with the photon spectrum .
figure [ rhospectrum ] compares the resulting spectrum with the input breit - wigner . with the breit - wigner ,
the cross section is about 5% lower than for a narrow resonance with the same coupling .
the phase space correction naturally cuts off the @xmath106 spectrum at @xmath113 ; we add an upper limit @xmath114 , matching the hera analysis from which the @xmath3 parameters are extracted@xcite .
the correction for higher masses is a few percent .
most studies of @xmath115 have modelled the exclusive @xmath116 spectrum with a breit - wigner @xmath4 plus a non - resonant @xmath116 term , and an interference term@xcite : @xmath117 the zeus collaboration found @xmath118 @xmath119b@xmath120 and @xmath121 ( @xmath119b / gev)@xmath120@xcite . with the interference term
, @xmath122 falls rapidly at large @xmath59 , and a correction for @xmath123 is negligible . as nuclear size increases , the non - resonant term becomes less important . at lower energies , for heavy nuclei it is at most a few percent . for large nuclei ,
the @xmath13 dependence of @xmath124 is unknwon , so we neglect non - resonant production . for the nuclei considered
, this should be less than a 10% effect on the @xmath116 rate .
the photonuclear cross sections are shown in figure [ sigmagamma ] .
the total cross sections follow a similar pattern to @xmath125 . at low energies ,
the nuclear form factor intrudes , and eliminates coherent production , with the cutoff energy depending on the nuclear radius and the final state mass .
the @xmath11 and @xmath6 curves cross because the @xmath11 includes a meson contribution that decreases with increasing energy , while the @xmath6 does not and because @xmath126 , with the exact relationship varying slightly with energy .
coherent photonuclear interactions are not a new concept .
several groups studied them experimentally and theoretically in the 1960s and 1970 s , using real photon beams .
more recent experiments , using virtual photons from muon scattering are less comparable because of the significant ( compared with @xmath20 ) photon virtuality .
most of the earlier calculations were used to extract parameters such as @xmath127 and @xmath21 from data .
our calculations are broadly similar . however , most of the earlier calculations used a disk model of the nucleus , with @xmath16 determined by a 2-dimensional fourier transform of a black disk , and @xmath15 contributing only to a phase shift .
many of these calculations also included a real part for the forward scattering amplitude . for the @xmath4 ,
the real part is 20 - 30% of the imaginary part , producing a 5% correction to the total amplitude@xcite . for other mesons
, the phases should be comparable ( when meson exchange is present ) or smaller , for pomeron - only interactions . in all cases ,
the correction should decrease slowly at higher energies .
the experimental results can also test our calculations .
because of the complications due to differing ( often poorly described ) treatments of backgrounds , @xmath4 width and the nonresonant @xmath116 , we limit our comparisons to narrow resonances . even for narrow resonances , there are complications .
few experiments cover the complete @xmath14 range ; extrapolations are required to find the cross sections .
we compare results using @xmath82 , because less extrapolation is required .
an experiment at cornell studied @xmath11 production from photons with average energy 8.2 gev striking a copper target , and found @xmath128 mb / gev@xmath129@xcite .
this compares well with our calculated @xmath130 mb / gev@xmath129 for that system .
another group studied @xmath6 production from 8.3 gev photons incident on copper and lead targets , and found @xmath131 and @xmath132 mb / gev@xmath129 respectively@xcite .
we predict higher cross sections , @xmath133 and 38 mb / gev@xmath129 respectively .
however , with the values of @xmath134 and @xmath135 used in the original analysis , we find @xmath136 and 26 mb / gev@xmath129 respectively , in agreement with the data . the discrepancy may stem from the large correction(60% for copper , 140% for lead ) to cover the full angular acceptance .
this correction is based on the optical model , and so depends on @xmath134 and @xmath135 .
still , these comparisons provide a useful check of our methods .
by integrating the photonuclear cross section over the photon spectrum , the total cross section is found : @xmath137 .
the corresponding production rates are shown in table [ ratevm ] .
these rates are very high , with exclusive meson cross sections about 10% of the total hadronic cross section for gold at rhic , and 50% for lead at lhc .
if the hadronic interaction model , eq . ( [ eqhint ] ) , is replaced with a hard sphere cutoff , ( @xmath138 ) , the rates rise about 5% . previously , we considered a model where @xmath139 cross sections were scaled from @xmath3 treating the nucleus as a black disk , with @xmath140 . in that model ,
the photonuclear cross sections had a different @xmath13 dependence , shown by the open triangles in fig .
[ dsdtgamma ] .
this scaling predicted cross sections much lower than are found here@xcite .
most other studies of photon - pomeron interactions have neglected the coherent pomeron - nucleus coupling .
baur , hencken and trautman found @xmath141=3.5 mb using the weizscker - williams photon flux for gold incident on a free proton target at rhic@xcite .
the final state rapidity is determined by the ratio of the photon energy and @xmath15 in the lab frame : @xmath142 so @xmath143 and @xmath144 is found by a change of variables . for the @xmath4 ,
the cross section must be integrated over @xmath106 .
@xmath145 is shown in fig .
[ dndy ] , for the @xmath4 ( breit - wigner ) , @xmath6 and @xmath2 , for production with gold at rhic and calcium at lhc .
the cross section is largest at low photon energies , corresponding to @xmath146 , because the photon flux drops as @xmath13 rises .
the peaking is larger for the @xmath4 and @xmath11 , where meson exchange increases the cross section at low energies . at rhic ,
when photons from both nuclei are added together , separate peaks appear in @xmath145 . at lhc ,
the higher energies spread the distribution over a much wider rapidity range , and the double peaked structure largely disappears , except for the @xmath4 .
the total perpendicular momentum spectrum is dominated by the two nuclear form factors , through the cutoff on the emitted photon @xmath147 and through the slope @xmath51 of @xmath94 .
the ( quadrature ) sum of these is roughly @xmath148 , or about 45 mev / c for gold .
this cutoff is a clear signature of coherent interactions , shared with @xmath149 and coherent double pomeron interactions . because of the exclusive production and small @xmath147
, these events can be easily separated from background such as grazing nuclear collisions , beam gas interactions and incoherent photonuclear interactions@xcite .
other authors have considered coherent double - pomeron interactions in heavy ion collisions@xcite .
if reactions where the nuclei collide are excluded , the cross section is very small , and depends critically on the range of the pomeron@xcite@xcite .
because the cross sections are so large , the probability of having multiple interactions in a single nucleus - nucleus encounter is nonnegligible .
this can be quantified by considering the meson production probability @xmath150 for a single collision .
figure [ probability ] shows this probability for various mesons for ( a ) gold beams at rhic and ( b ) lead beams at lhc ; the probabilities reach the 1% level . for the @xmath4
, we show results from two different hadronic interactions models : the woods - saxon model , eq .
( [ eqhint ] ) , and also a hard sphere @xmath138 result .
although the two models predict very similar overall rates , the probability is significantly affected around @xmath151 , doubling the maximum probability at lhc .
this probability is high enough that , even in the absence of correlations@xcite , significant numbers of vector meson pairs should be produced .
neglecting correlations , the pair production probability is @xmath152 for identical pairs and @xmath153 for nonidentical pairs .
these cross section and corresponding yearly rates are tabulated in table [ sigmapair ] .
triples and higher multiples are also possible .
of course , final state interactions may affect what is observed in a detector .
two - photon physics is expected to be a significant activity at rhic @xcite@xcite and lhc@xcite@xcite . because the kinematics for both photonuclear interactions and two - photon collisions is dominated by the coherent coupling to the two nuclei , and consequently by the nuclear form factors , the reactions have similar kinematics .
so , the specific pathway may not be determinable from the final state
. the rates for photonuclear interactions are considerably higher than for comparable two - photon reactions .
for example , the @xmath6 production rate with gold beams at rhic is about 8 hz . for comparison ,
the rates for @xmath154 and @xmath155@xcite are 0.13 and 0.02 hz respectively . at higher masses ,
the situation is similar , and photonuclear vector meson production may overshadow two - photon production of scalar and tensor final states , complicating meson spectroscopy .
it is worth noting that the photonuclear @xmath156 rate is many times larger than that expected from @xmath149 production@xcite .
on the other hand , the same techniques developed to study @xmath149 physics at heavy ion colliders are effective at selecting photonuclear interactions .
the similarity of the two interactions can lead to some interesting opportunities .
for example , when the final states are indistinguishable , interference can occur between the photonuclear and @xmath149 production channels .
interference between leptonic decays of photonuclear produced vector mesons and bethe - heitler @xmath157 pairs has been studied previously@xcite ; at future heavy ion colliders , many more channels will be accessible . despite the similarity
, the different @xmath145 spectra may allow statistical separation of @xmath149 and photonuclear interactions .
this is because the photon flux scales as @xmath158 , while the pomeron / meson couplings have a much weaker energy dependence .
the @xmath145 distribution for photonuclear production is broader than for two - photon production .
this is especially true for photon - meson interactions , where the @xmath158 photon spectrum combines with a cross section that drops with energy to form two peaks at large positive and negative rapidity , corresponding to low energy photons from the two nuclei in turn .
figure [ dndycompare ] compares the two distributions for @xmath159 and for production of a hypothetical vector meson with the same mass .
two vector meson models are shown : one based on the @xmath11 couplings and another with only pomeron coupling .
the two are very similar at @xmath160 , but the @xmath11 model peaks about 10% higher
. it may also be possible to separate the two classes by their different impact parameter dependence .
a two - photon interaction can occur at a significant distance from both nuclei , while a photonuclear interaction must occur inside or very near a nucleus .
so , the two classes will have a different impact parameter dependence . in a collision
, photon exchange can leave the nuclei excited in a giant dipole resonance , regardless of other activity ; the excitation probability depends on the impact parameter@xcite .
this may allow us to statistically separate the two classes . in a broader study
, it will also be possible to separate the two classes by measuring how the rates vary with @xmath8 ; the meson and pomeron flux rises should be less @xmath8 dependent than photons .
this study will require data with a variety of different beam species .
we have calculated the rates and rapidity distribution for exclusive photonuclear production of vector mesons in heavy ion collisions , using a glauber model calculation .
the lhc results involve photons with energies considerably above those currently available .
although this adds to the physics interest , it also introduces some uncertainty into the rates presented here . for the @xmath2 ,
in particular , the spectrum must soften at some photon energy , and the rates given here may be overestimates .
the production rates are large enough that heavy ion colliders could be used as vector meson factories .
the @xmath6 and @xmath2 production rates at lhc are comparable to those at existing or planned meson factories based on @xmath157 annihilation .
these rates are higher than comparable two - photon or double - pomeron interactions .
thus , these photonuclear interactions should be the dominant source of exclusive interactions at heavy ion colliders .
in addition to single interactions , multiple vector meson production from the same ion pair should be measurable .
we thank jrgen randrup for suggesting the hard sphere plus yukawa potential form factor .
we would like to acknowledge useful conversations with stan brodsky and ralph engel .
this work was supported by the us doe , under contract de - ac-03 - 76sf00098 .
.luminosity and beam kinetic energy for several heavy ion beams at rhic and lhc . because rhic will be dedicated to heavy ion acceleration , it is likely to run a wider variety of beams than lhc .
the rhic luminosities are from ref .
different references quote somewhat different heavy ion luminosity for lhc ; these numbers are calculated from table 3 of ref .
@xcite , assuming 2 experiments , and that ( as suggested ) average luminosity is 45% of peak luminosity . [ cols="<,>,>,>",options="header " , ] | exclusive vector meson production reactions such as au + au @xmath0 au + au + v , where @xmath1 or @xmath2 can proceed through photon - pomeron and photon - meson interactions .
photons from the electromagnetic field of one nucleus interact coherently with the other nucleus .
photonuclear cross sections are scaled from @xmath3 data , and convoluted with the photon spectrum to find the exclusive rates .
the cross sections at the rhic and lhc heavy ion colliders are huge , 10% of the total hadronic cross section at rhic , and 50% at lhc
. these accelerators may be useful as vector meson factories . with iodine beams at rhic
, 640 @xmath4 are produced each second ( @xmath5/year ) ; with calcium at the lhc the rate is 240 khz .
the @xmath6 rates are 39 hz at rhic and 15 khz at lhc , while the @xmath2 rate is 0.3 hz at rhic and 780 hz at the lhc .
because of the coherent couplings , the reactions kinematics are similar to coherent two - photon interactions ; we discuss the interplay between the two reactions .
( submitted to physical review c. ) |
the top quark was discovered at the tevatron in 1995 @xcite . since then , many of its properties have been measured at the tevatron and the large hadron collider ( lhc ) .
almost all of these seem to agree rather well with the standard model @xcite . for a while ,
the forward - backward asymmetry in @xmath1 production at the tevatron seemed to be much larger than predicted in the standard model @xcite .
however , in recent analyses , the discrepancy seems to have decreased considerably @xcite .
moreover , at the lhc , no deviation was observed at any stage in the analogous charge asymmetry observables @xcite . at present , the top quark does seem to behave exactly as described in the standard model .
nonetheless , the lhc , being a top - factory , affords us the opportunity to probe top properties further and also study rare processes involving the top .
one such process is the decay @xmath0 .
the dominant decay modes of the top involve @xmath2 , with the @xmath3 then decaying to @xmath4 , @xmath5 or @xmath6 .
the decay @xmath0 would arise if @xmath7 . however , this decay is extremely suppressed in the standard model as it is proportional to @xmath8 ( @xmath9 = 0.04 being one of the elements of the cabibbo - kobayashi - maskawa matrix ) . new physics contributions to this decay
, if present , can be parametrized in terms of the following effective lagrangian : @xmath10 where @xmath11 { \cal l}_{eff}^t \ ; & = \ ; 4\sqrt{2}g_f v_{cb}v_{tb } \big\ { x^t_{ll } \ ; ( \bar b \sigma^{\mu\nu } p_l t ) \ ; ( \bar c \sigma_{\mu\nu } p_l b ) \ ; + \ ; x^t_{rr } \ ; ( \bar b \sigma^{\mu\nu } p_r t ) \ ; ( \bar c \sigma_{\mu\nu } p_r b ) \ , \big \ } \ ; + \ ; h.c.\end{aligned}\ ] ] in the above , the colour indices have been suppressed but they are assumed to contract in the same manner as in the standard model . at the lhc , the largest number of top quarks are produced through the process @xmath12 .
the decays of these tops and antitops can then be investigated for signs of new physics .
the large production cross - section compensates for the small branching fraction such that a statistical analysis remains feasible .
moreover , as shown in refs .
@xcite and @xcite , angular correlations between @xmath13 and @xmath14 decay products also carry the imprint of new physics .
for the purposes of this analysis , it is assumed that the @xmath13 decays to @xmath15 , @xmath16 and @xmath17 , while the @xmath14 decays to a @xmath16 , a charged lepton and an antineutrino .
the square of the matrix element for the process @xmath18 is calculated in ref .
it is seen that all new - physics dependence in @xmath19 can be parametrized in terms of certain combinations @xmath20 of the couplings @xmath21 given by @xmath22 % \hat{a}_{\bar b}^- \quad & = \quad 4\left|x^{v}_{rr}\right|^2 - 8 \,\mbox{re}\left(x^t_{rr}x^{s*}_{rr}\right)+32 \left|x^t_{rr}\right|^2 \ , , \\[0.5ex ] % \hat{a}_{b}^+ \quad & = \quad \left|x^{s}_{ll}\right|^2+\left|x^{s}_{lr}\right|^2 -16\left|x^{t}_{ll}\right|^2 \ , , \\[0.5ex ] % \hat{a}_{b}^- \quad & = \quad \left|x^{s}_{rr}\right|^2+\left|x^{s}_{rl}\right|^2 -16\left|x^{t}_{rr}\right|^2 \ , , \\[0.5ex ] % \hat{a}_{c}^+ \quad & = \quad 4\left|x^{v}_{lr}\right|^2 + 8 \,\mbox{re}\left(x^t_{ll}x^{s*}_{ll}\right)+32\left|x^t_{ll}\right|^2 \ , , \\[0.5ex ] % \hat{a}_{c}^- \quad & = \quad 4\left|x^{v}_{rl}\right|^2 + 8 \,\mbox{re}\left(x^t_{rr}x^{s*}_{rr}\right)+32\left|x^t_{rr}\right|^2 \ , .\end{aligned}\ ] ] in fact , as discussed in ref .
@xcite ( sec .
iii b 3 ) , this is true even for @xmath23 .
if the @xmath13 decay contains contributions from new physics , then , several observables related to the decay products get affected .
of these , two turn out to be specially useful .
these are 1 .
@xmath24 where @xmath25 and @xmath26 and @xmath27 are the momenta of two out of the three particles coming from the top decay .
@xmath28 where @xmath29 is the angle between the direction of the @xmath30 decay product of the @xmath13 and the spin quantization axis of the @xmath13 in the @xmath13 rest frame , and @xmath31 is the angle between the direction of the @xmath32 and the spin quantization axis of the @xmath14 in the @xmath14 rest frame .
the detailed expressions for these for different choices of @xmath33 and @xmath34 can be found in ref .
[ fig : observables ] shows the normalised distributions for @xmath35 and @xmath36 for a particular choice of new - physics couplings .
it clearly demonstrates the discriminating power of these observables even for cases where the new - physics couplings are not large .
that is not all . @xmath37 and @xmath38 form a set of six differential distributions , each dependent on a different combination , there is also some additional dependence on re(@xmath39 ) . ]
of the six @xmath20 s . by performing a simultaneous fit to all six distributions ,
it is possible to extract the values of these @xmath20 s .
@xcite discusses the detailed procedure for carrying out this exercise . in the absence of actual data in this channel ,
monte carlo data generated using madgraph 5 @xcite is used to test and establish the algorithm .
the model files for madgraph have , in turn , been generated using feynrules @xcite .
the crux of the idea is as follows .
each of @xmath37 and @xmath38 gives a histogram .
in general , each histogram @xmath40 can be written as @xmath41 where @xmath42 is the histogram @xmath40 for the standard model and other @xmath43 s represent same histogram @xmath40 with @xmath20 and re(@xmath39 ) = @xmath44 , one at a time .
in essence , one only needs to determine the weights @xmath45 that best reproduce the data .
the distributions @xmath38 are , in fact , closely related to the @xmath46 spin correlations . in the standard model , @xmath47 where @xmath48 is the spin correlation coefficient and the factors @xmath49 and
@xmath50 depend on the identity of the particles @xmath33 and @xmath51 .
since the @xmath1 pair decays instantly , the spin correlation coefficient is inferred by fitting this distribution to the functional form @xmath52 .
the coefficient @xmath53 contains @xmath48 . in the presence of new physics ,
this distribution gets modified to the form @xmath54 where @xmath55 is a different function of the @xmath20 s for @xmath56 .
if the value of the spin correlation coefficient were to be inferred from this now , it would yield @xmath57 .
there is , however , a caveat in all of this .
note that the final state contains two @xmath16 s . in order to construct the aforementioned observables , the @xmath16 coming from the @xmath13-decay has to be ` correctly ' identified . here and in ref .
@xcite , the @xmath16 that yields the smaller value of @xmath58 is taken to be the one arising from the @xmath13-decay .
if both @xmath16 s yield nearly identical values of @xmath59 , the event is discarded .
this ` cut ' modifies the shape of the angular correlations such that it is no longer possible to fit them to the functional form @xmath52 or any simple modification of it .
nonetheless , the histogram fitting described in the previous section works just as well since the same ` cut ' is applied to the data ( @xmath40 ) as well as the templates ( @xmath43 ) .
a full fit to all six observables is performed using monte carlo data with statistics corresponding to an integrated luminosity @xmath60 300 fb@xmath61 for 14 tev collisions at the lhc .
different new - physics scenarios corresponding to different choices of the @xmath21 s are examined .
the values of @xmath20 are recovered to within @xmath62 1.7@xmath63 of their input values .
if only the @xmath64 distributions are measured , the values of ( @xmath65 ) , ( @xmath66 ) and ( @xmath67 ) can still be determined .
99 t. aaltonen _ et al . _ [ cdf collaboration ] , phys .
d * 87 * , 092002 ( 2013 ) . v. m. abazov _ et al .
_ [ d0 collaboration ] , phys . rev .
d * 90 * , 072011 ( 2014 ) .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1402 * , 107 ( 2014 ) s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1404 * , 191 ( 2014 ) .
k. kiers , p. saha , a. szynkman , d. london , s. judge and j. melendez , phys .
d * 90 * , 094015 ( 2014 ) .
p. saha , k. kiers , d. london and a. szynkman , phys .
d * 90 * , 094016 ( 2014 ) .
j. alwall _ et al .
_ , jhep * 1106 * , 128 ( 2011 ) , http://madgraph.hep.uiuc.edu . | top physics provides a fertile ground for new - physics searches . at present
, most top observables appear to be in good agreement with the respective standard model predictions .
however , in the case of decay modes that are suppressed in the standard model , new - physics contributions of comparable magnitude may exist and yet go unnoticed because their impact on the total decay width is small .
hence it is interesting to probe rare top decays .
this analysis focuses on the decay @xmath0 .
useful observables are identified and prospects for measuring new - physics parameters are examined . |
the x - ray emission from radio - quiet active galactic nuclei ( agn ) shows persistent variability that is stronger and more rapid than in any other waveband .
this rapid x - ray variability provided early support for the black hole / accretion disc model of agn ( e.g. rees 1984 ) and suggested that the x - ray emission is produced in the inner regions of the central engine ( see mushotzky , done & pounds 1993 for a review ) .
the x - ray fluctuations appear erratic and random , with variability occurring on a range of timescales ( e.g. m@xmath3hardy 1989 ) . in the study of time - variable processes one of the most widely used statistical tools
is the variability power spectrum ( e.g. priestley 1981 ; bloomfield 2000 ) .
the power spectral density ( psd ) describes the amount of variability power ( i.e. amplitude@xmath7 ) as a function of temporal frequency ( timescale@xmath8 ) .
the long , uninterrupted observations provided by allowed for the measurement of the first accurate x - ray psds for seyfert 1 galaxies . these psds were shown to rise at lower frequencies as a power - law : @xmath9 , where @xmath10 is the power at frequency @xmath11 and @xmath12 is the psd slope , found to be @xmath13 over the frequencies @xmath14 to @xmath15 hz ( green , m@xmath3hardy & lehto 1993 ; lawrence & papadakis 1993 ) .
this type of broad - band variability , rising to lower frequencies with a power spectral slope @xmath16 , is usually called `` red noise '' ( for an introduction to red noise in astronomy and elsewhere see press 1978 ) .
the red noise psds of seyfert 1s are similar to those observed in galactic black hole candidates ( gbhcs , see e.g. van der klis 1995 ) on much shorter timescales , and suggests the x - ray emission mechanisms may be the same in these objects that differ in black hole mass by factors of @xmath17 ( m@xmath3hardy 1989 ) .
the psd of the best studied gbhc , cygnus x-1 in its low / hard state , can be approximated by a doubly - broken power - law with a steep slope at high frequencies ( @xmath18 above @xmath19 hz ) , breaking to a flatter slope at intermediate frequencies ( @xmath1 in the range @xmath20 hz ) then breaking to a flat ( `` white '' ) spectrum at low frequencies ( @xmath21 below @xmath22 hz ) .
see nowak ( 1999a ) and belloni & hasinger ( 1990 ) for more details of cyg x-1 .
other gbhcs such as gx 3394 show similar behaviour ( nowak , wilms & dove 2002 ) .
the position of the breaks in the psd represent `` characteristic timescales '' of the system , but these timescales have no clear physical interpretation at present ( e.g. belloni , psaltis & van der klis 2002 and references therein ) .
the featureless psds of seyfert 1s provided no such timescales .
such breaks must exist , however , or the total power would diverge at low frequencies ( for psd slopes @xmath23 ) .
indeed , in recent years dedicated x - ray monitoring programmes with have revealed deviations from the power - law psd in a few seyfert 1 galaxies ; at low frequencies the psd breaks to a flatter slope ( edelson & nandra 1999 ; uttley , m@xmath3hardy & papadakis 2002 ; markowitz 2002 ) .
the timescale of this break frequency is consistent with the idea that the psd of agn and gbhcs are essentially the same , with the characteristic timescales scaling linearly with the black hole mass .
such a linear scaling of timescales with black hole mass might be expected if the characteristic timescales are related to e.g. a light - crossing timescale or an orbital timescale .
this linear scaling also means that a 100 ks light curve of a seyfert galaxy ( such as presented here ) is comparable to a @xmath24 s light curve of cyg x-1 . on these timescales
a bright seyfert 1 galaxy provides many more photons ( by more than two orders of magnitude ) than the gbhc , i.e. seyferts have a higher count rate per characteristic timescale .
this means that intensive monitoring of seyferts can provide a view of the very high frequency variations in accreting black holes that is difficult to access even in bright gbhcs .
not only do agn and gbhcs show similar psds , but they display other striking similarities in their x - ray timing properties .
uttley & m@xmath3hardy ( 2001 ) showed that both x - ray binary and agn light curves show a tight , linear correlation between flux and rms variability amplitude ( see also edelson 2002 ) .
the variability properties of gbhcs such as cyg x-1 are a function of photon energy .
the psd shape above the high frequency break is energy - dependent , with the harder energy bands showing flatter psd slopes ( e.g. nowak 1999a ; lin 2000 ) .
in addition the variations between different bands are highly coherent ( well correlated at each fourier frequency ) and show time delays , with variations at softer energies leading those at harder energies ( miyamoto & kitamoto 1989 ; cui 1997 ; nowak 1999a ) .
the magnitude of the time delay is frequency dependent , with the delay increasing at lower frequencies . to date these properties
have only been measured in one agn , ngc 7469 ( nandra & papadakis 2001 ; papadakis , nandra & kazanas 2001 ) , and in this object the energy - dependence of the psd slope , the coherence and time delay properties appeared remarkably similar to those seen in cyg x-1 .
these similarities re - enforce the idea that the x - ray emission mechanisms are the same in gbhcs and agn .
this paper presents a detailed analysis of the x - ray continuum variability of the bright , variable seyfert 1 galaxy ( @xmath25 ) using a long observation .
the large collecting area , wide band - pass and long orbit ( @xmath26 day ) of make it ideal for examining the high frequency ( @xmath27 hz ) timing properties of seyfert galaxies .
the analysis concentrates on frequency domain ( i.e. fourier ) time series analysis techniques , as these are used extensively in the analysis of gbhc data .
the rest of this paper is organised as follows . in section [
sect : data ] the data reduction procedures are outlined . in section [
sect : timing ] the basic properties of the light curves are discussed , followed by an analysis of the psd in section [ sect : pds ] .
spectral variability is then examined in detail by measuring the coherence and time lags between different energy bands , as discussed in section [ sect : cross ] . as these analyses are relatively new to agn studies we discuss the methodology in some detail .
finally , the implications of these results are discussed in section [ sect : disco ] and some concluding remarks are given in section [ sect : conc ] .
( jansen 2001 ) observed over the period 2001 july 31 2001 august 5 ( revolutions 301 , 302 and 303 ) , during which all instruments were operating nominally .
fabian ( 2002 ) discuss details of the observation and basic data reduction . for the present paper
the data were reprocessed entirely with the latest software ( sas v5.3.3 ) but the procedure followed that discussed in the previous paper ( except that the extraction region was a circle of radius 35 arcsec for both mos and pn ) .
light curves were extracted from the epic pn data in four different energy bands : 0.210.0 kev ( full band ) , 0.2 - 0.7 kev ( soft band ) , 0.7 - 2.0 kev ( medium band ) and 2.0 - 10.0 kev ( hard band ) . these were corrected for telemetry drop outs ( less than 1 per cent of the total time ) , background subtracted and binned to 100 s time resolution .
the errors on the light curves were calculated by propagating the poisson noise .
the light curves were not corrected for the @xmath28 per cent `` live time '' of the pn camera ( strder 2001 ) , which is only a scaling factor .
the epic mos light curves were also examined , but these are essentially identical to the pn light curves except with lower signal - to - noise .
therefore , in the rest of this paper the analysis concentrates on the pn data , but the mos data are used to confirm the pn results of section [ sect : cross ] .
the full band light curves are shown in fig .
[ fig : lightcurves ] .
the revolutions provided three uninterrupted light curves of 80 , 123 and 124 ks duration .
as discussed in fabian ( 2002 ) , observed simultaneously with , but these data are not used in the present paper .
the lecs and mecs data cover approximately the same bandpass as the epic ( but with lower signal to noise ) and the higher energy pds data ( with a count rate @xmath29 ct s@xmath8 ) do not have sufficient statistics to determine the high energy variability properties accurately .
in addition , analysis of the data is complicated by the poor sampling of the light curves ( due to repeated earth occulatations and saa passages ) , indeed the amount of `` good '' exposure time obtained for the pds was @xmath30 ksec spread over a @xmath31 ksec observing period .
the source shows a factor of five variation between minimum and maximum flux ( both occurring during revolution 303 ) , and significant variability can be seen on timescales as short as @xmath32 s. as a first measure of the variability amplitude the root mean square ( rms ) amplitude , in excess of poisson noise , was calculated as @xmath33 where @xmath34 is the total variance of the light curve @xmath35 of length @xmath36 data points @xmath37 where @xmath38 is the mean of @xmath35 and @xmath39 is the contribution expected from measurement errors @xmath40^{2},\ ] ] and @xmath41 is the error on @xmath35 .
the fractional excess rms variability amplitude ( @xmath42 ; edelson 2002 ) and mean count rate ( @xmath38 ) were calculated for each of the three revolutions and are tabulated in table [ tab : revs ] .
the errors on @xmath43 were estimated using the prescription given by edelson ( 2002 ; but see their appendix a for caveats on the interpretation of these estimates ) .
the medium band showed the highest variability amplitude .
this can also been seen in fig .
6 of fabian ( 2002 ) , where the fractional rms spectrum peaked in the 0.72.0 kev band . .
mean count rates ( @xmath38 ) and fractional excess rms variability amplitudes ( @xmath43 ) for the light curves from each revolution and in each energy band .
[ tab : revs ] [ cols="<,^,^,^",options="header " , ] the binned periodograms used above reached down to @xmath44 hz in the lowest frequency bin .
it would in principle be possible to treat the three consecutive light curves as a single , longer light curve ( top panel of fig . [
fig : lightcurves ] ) and use this to reach even lower frequencies in the power spectrum .
however , this three revolution light curve is no longer uninterrupted , due to the gaps between revolutions .
this means that the standard dft ( as discussed in section [ sect : periodogram ] ) is no longer suitable .
there do exist periodogram estimators designed to deal with data containing gaps , such as the lomb - scargle periodogram ( lomb 1976 ; scargle 1982 ) but there are drawbacks to using such techniques .
firstly , the gaps in the sampling mean the periodogram estimates at each frequency are no longer independent of one another ( scargle 1982 ) , so the binning and error estimation used above is not strictly valid .
secondly , the more complicated window function introduces much greater distortion on the low frequency periodogram , which would severely limit the amount of information that could be recovered from the lowest frequencies . for these reasons
the results presented in this paper are based entirely on the more robust , standard dft method , but for completeness the lomb - scargle periodogram ( computed using the algorithm of press & rybicki 1989 ) of the full - band light curve is shown in fig .
[ fig : lomb ] .
down to low frequencies this appears broadly consistent with the results presented above .
the fundamental assumption underpinning the above analysis is that the variability process is stationary .
a stationary process is one whose statistical properties do not depend on time ( see section 1.3 of bendat & piersol 1986 for more on stationarity in random data ) .
specifically , the assumption made here is that the psd is stationary , i.e. that the light curves are all realisations of the same process .
the rms - flux correlation found in section [ sect : rms - flux ] shows that the variability of is in some sense non - stationary ( the variance does change ) , but in a `` well - behaved '' fashion .
indeed , the linear correlation between @xmath45 and @xmath38 means that the average @xmath43 is constant . in other words ,
the variations , when normalised by the local flux , appear stationary .
the method outlined in appendix a of papadakis & lawrence ( 1995 ) was used as a simple check of the validity of this assumption .
if the psd @xmath10 is constant then the ( unbinned ) periodograms obtained from each revolution should be identical except for the scatter dictated by equation [ eqn : pds_scatter ] .
papadakis & lawrence ( 1995 ) define the statistic @xmath46 from the logarithm of the ratio of two periodograms , summed over a range of frequencies .
@xmath47-\log[p_2(f_i)]}{\sqrt{{\rm var}\{\log[p_1(f_i)]\}+{\rm var}\{\log[p_2(f_i)]\ } } } , \ ] ] where @xmath48 and @xmath49 are the two periodograms , and @xmath50\}=0.31 $ ] ( papadakis & lawrence 1993 ) .
if the underlying psd is the same in the two periodograms then @xmath46 should be normally distributed with a mean of zero and a variance of unity .
a normalised periodogram was computed for the first 80 ksec of each of the three revolutions ( the light curves were clipped to equal duration so that their periodograms had identical fourier frequencies ) and @xmath46 was measured for frequencies below @xmath51 hz , where the source variability dominates over the poisson noise . comparing rev .
301 and 302 gave @xmath52 , comparing rev .
301 and 303 gave @xmath53 and comparing rev .
302 and 303 gave @xmath54 .
these three estimates are all within @xmath55 of the expected value assuming stationarity in the data .
therefore , to first - order at least , the assumption of stationarity in these data seems a reasonable one .
the psd is often called the auto spectrum because it is the fourier transform of the light curve , multiplied by itself ( equation [ eqn : ft2 ] ) .
the cross spectrum is a related tool used for comparing the properties of two simultaneous light curves , such as from different energy bands , say @xmath56 and @xmath57 .
the product of the fourier transforms of the two light curves gives the cross spectrum ( compare with equation [ eqn : ft2 ] ) : @xmath58 in the same way that the psd ( auto spectrum ) and the acf are fourier pairs , the cross spectrum and cross correlation function ( ccf ) are also fourier pairs .
thus , mathematically speaking , the cross spectrum contains the same information as the ccf .
the cross spectrum is complex valued and can be represented in the complex plane by an amplitude and a phase .
the complex fourier transforms of the two light curves can be written as @xmath59 and @xmath60 and the cross spectrum can be written as : @xmath61 the phase component of the cross spectrum represents the phase difference between the two light curves ( which corresponds to a time delay at that frequency ) and is discussed more fully in section [ sect : phase ] .
the squared magnitude of the cross spectrum is used to define the `` coherence '' of the two light curves .
the coherence , @xmath62 , is a real - valued function and is essentially the squared cross spectrum normalised by the auto spectra ( psds ) of the two light curves .
@xmath63 here the angled brackets represent an averaging over an ensemble of realisations of the processes .
coherence differs from the periodogram and time lags in the sense that it is only meaningful to talk of the coherence from an ensemble of independent measurements ( either from separate light curve segments or consecutive frequencies ) .
equation [ eqn : coherence1 ] defines the linear correlation coefficient of @xmath64 and @xmath65 ( see section 9.1 of priestley 1981 ) defined by equation [ eqn : coherence1 ] is often referred to as the _ squared coherence _ ( priestley 1981 ; bendat & piersol 1986 ) and it gives the square of the linear correlation coefficient .
this function is simply referred to as the coherence in this paper . ] .
the coherence can thus be thought of as a measure of the degree of linear correlation between the two light curves as a function of fourier frequency , and is related to the amplitude of the ccf . in order to illustrate coherence , consider two light curves that are related by a transfer function @xmath66 , e.g. @xmath67 then from the convolution theorem @xmath68 where @xmath69 is the fourier transform of @xmath70 .
therefore : @xmath71 and @xmath72 the squared magnitude of the cross spectrum then becomes : @xmath73 therefore the coherence is @xmath74 .
this means that if the two light curves are related by a single , linear transfer function ( e.g. a delay or a smoothing ) then they will have unity coherence . in other words ,
if knowledge of one light curve can in principle be used to predict the other , the two are said to be perfectly coherent . if , on the other hand , the light curves contain contributions from completely independent processes then the coherence will be less than unity .
another way of thinking about coherence is in terms of equation [ eqn : cross_spec2 ] .
if the light curves are related by some simple transform then , at a frequency @xmath11 , there will be a fixed delay between the light curves , i.e. the lag will be the same in each realisation of the process . if the phase difference is constant then the value of the numerator of equation [ eqn : coherence1 ] ( the squared mean of equation [ eqn : cross_spec2 ] ) is equal to the denominator and hence the the coherence is unity .
the interpretation of coherence is illustrated more fully in vaughan & nowak ( 1997 ) and nowak ( 1999a ) . in an analogous manner to the psd
, the cross spectrum is estimated using the cross periodogram .
if the two discretely sampled light curves , @xmath75 and @xmath76 , have dfts @xmath77 and @xmath78 then the cross periodogram is given by the discrete form of equation [ eqn : cross_spec ] , namely : @xmath79 where @xmath80 and @xmath81 are the real and imaginary components of the complex cross spectrum : @xmath82 and @xmath83 and @xmath84 are the real and imaginary components of the dft of the light curve @xmath75 as defined by equation [ eqn : sine ] .
the coherence is estimated using the discrete form of equation [ eqn : coherence1 ] : @xmath85 the angled brackets here represent an averaging over light curve segments and/or consecutive frequencies . in practice
however , the coherence will be compromised by photon noise , which is independent between the two light curves and so will lead to an apparently reduced coherence .
as the poisson noise power is known , this effect can be removed using the recipe outlined in vaughan & nowak ( 1997 ) . in the following analysis
the noise - corrected coherence and its uncertainty were estimated using equation 8 of vaughan & nowak ( 1997 ) .
the coherence was calculated only for frequencies below @xmath51 hz , above this frequency the poisson noise power becomes comparable to the intrinsic source variability and equation 8 of vaughan & nowak ( 1997 ) becomes increasingly inaccurate . as mentioned above , the coherence needs to be averaged in some fashion . for the present analysis
the cross and auto spectra were calculated separately for each revolution .
the data from the three revolutions were then combined by sorting in order of frequency and then averaged in bins of @xmath86 frequency points .
the coherence function was calculated for the three possible combinations of the three energy bands ( soft / medium , soft / hard and medium / hard ) .
this was repeated using data from mos1 and mos2 as well as the pn and the results were consistent between the detectors .
( the coherence was also estimated individually for each revolution of data and the results were consistent between the three revolutions . )
the resulting coherence functions are shown in fig . [
fig : coherence1 ] . at low frequencies
the coherence is close to unity for all energy bands . at higher frequencies
the coherence becomes much lower , particularly between the soft and hard bands . for comparison , fig .
[ fig : coherence1 ] shows the function @xmath87 compared to the coherence between soft and hard bands .
this apparent loss of coherence at high frequencies could be an effect of poisson noise in the data , since at high frequencies the intrinsic source power is weaker compared to the noise power . in order to test this possibility
the coherence was evaluated by comparing light curves in identical energy ranges from the three different detectors .
the light curves extracted from the three detectors should be identical , in a given energy band , except for poisson noise .
therefore the coherence should be unity and any deviation from unity will be a result of poisson noise .
as can been seen from fig .
[ fig : coherence2 ] , the coherence is close to unity until the highest frequencies , and the deviation from unity coherence at high frequencies is much smaller than the observed loss of coherence between the different energy bands .
this , and the fact that the loss of coherence is repeated in all three detectors suggests that poisson noise is not seriously biasing these results , and that the error bars are reasonable estimates of the true uncertainties .
simulated data were used as an additional test of the accuracy of the coherence estimation .
two simulated light curves were constructed , representing light curves in two different bands , and the coherence was calculated for them exactly as for the real data .
these simulations were used as a further test of the effect of poisson noise and to measure the effect of different psd slopes on the coherence function .
the effect of poisson noise was tested by simulating coherent data in two bands using the same psd .
the psd model used was taken from the fit to the soft band data using model 2 ( table [ tab : pds_fit2 ] ) .
two long light curves were generated using this psd and an identical random number sequence .
the two light curves , having identical psds as well as fourier phases and amplitudes , were thus identical and therefore have unity coherence at all frequencies .
the simulated light curves for the two bands were resampled to match the window function of the real data , i.e. three consecutive light curves were extracted from each to band match the three revolutions of data .
the light curves for the first band were rescaled to match the average count rate of the soft band light curve and the light curves from the second band were rescaled to match the count rate of the hard band data .
( the simulated light curves were produced using the absolute psd normalisations found from fitting the data .
thus the absolute variability amplitude of the rescaled simulated data is as expected based on the real data . )
poisson noise was then added independently to the simulated light curves ( i.e. the number of counts in each time bin was randomised according to the poisson distribution ) .
these two sets of simulated data represent soft and hard band light curves in terms of their poisson noise , but intrinsically are perfectly coherent and have the same psd .
one hundred sets of simulated data were produced and for each set the coherence and its error were calculated as discussed in section [ sect : coh_est ] ( i.e. using the recipe of vaughan & nowak 1997 ) . the average and standard deviation of the coherence estimates from the 100 sets of simulated data
are shown in the top panel of fig .
[ fig : coherence_sim ] . also shown
are the average error bars .
the coherence is clearly very close to unity over the frequency range examined , and the standard deviation of the estimates is close to the mean error bar , indicating that the vaughan & nowak ( 1997 ) method did accurately recover the intrinsic coherence from the effect of the poisson noise .
( note that these error bars are slightly smaller than those derived from the real data because the uncertainty is itself a function of the intrinsic coherence , which is lower in the real data . )
the tests above used intrinsically coherent data with identical psd shapes .
the psds fits of section [ sect : e - psd ] show that the psd shape is energy dependent . in order to confirm that the coherence is not biased because of the psd shape being different in different energy bands
, the experiment was repeated using psds appropriate for the two bands .
specifically , the best - fitting psd models for the soft and hard bands ( listed in table [ tab : pds_fit2 ] ) were used to generate simulated soft and hard bands light curves , but again the two sets of data used an identical random number sequence .
thus the simulated soft and hard band light curves again have unity intrinsic coherence but different psds .
as before , 100 sets of simulated data were used and the average coherence and error were calculated .
the results are shown in the bottom panel of fig .
[ fig : coherence_sim ] .
the average coherence is again very close to unity , indicating the different psd slopes in the different bands did not bias the results .
the loss of coherence observed in is thus not an artifact of poisson noise in the data , nor the change in psd slope with energy , and is repeated in each of the three epic cameras . as such , in the discussion below , it is assumed that this loss of coherence is intrinsic to the continuum variations of the source .
as mentioned in section [ sect : cross spectrum ] , the cross spectrum also contains information about the phase differences between the two light curves ( equation [ eqn : cross_spec2 ] ) .
this information is similar to the delays measured by the ccf , except measures the delay as a function of fourier frequency .
the phase lag is measured from the phase of the averaged cross periodogram : @xmath88 where @xmath89 is calculated using equation [ eqn : cross_spec3 ] and the angled brackets denote an averaging over consecutive frequency points and/or light curves .
this phase difference can then be converted into a time delay at fourier frequency @xmath11 : @xmath90 only if the coherence is high at a given frequency is the time lag meaningful . if the two light curves are not coherent then the individual phase estimates from the two light curves are poorly correlated and so the phase difference of the average cross spectrum has no useful interpretation . for the present analysis , the time lag was calculated using equations [ eqn : phase1 ] and [ eqn : time_lag ] , with the averaging of the cross spectrum performed in bins of @xmath86 frequency points .
the error on the time lag estimate was computed using equation 16 of nowak ( 1999a ) .
see section 9.1.3 of bendat & piersol ( 1986 ) for a derivation of this equation .
the effect of poisson noise is to add an independent , random component to the phases of the two light curves and hence randomise the phase difference .
this means that small time lags are difficult to detect due to photon noise .
the recipe outlined in nowak ( 1999a ) ( see their section 4.2 ) was used to estimate the effective noise limit , below which it is difficult to detect reliable lags . the estimated error on the phase lag
is computed using their equation 14 : @xmath91 this uncertainty was calculated using values appropriate for the hard band ( psd normalisation , slope and noise level @xmath92 from section [ sect : pds_results ] , and coherence estimated using the function given in section [ sect : coherence_results ] ) .
this estimated uncertainty , which should be considered merely a lower limit on uncertainty introduced by poisson noise , was compared to the function used in section [ sect : phase_results ] to illustrate the observed time lags and the results are shown in fig .
[ fig : sensitivity ] .
this indicates that above @xmath93 hz the lags are too small to be detected , i.e. they fall below the sensitivity limit defined by the poisson noise .
a further point is that the coherence is significantly below unity at higher frequencies ( @xmath94few @xmath95 hz ) , in which case any measured time lag is no longer meaningful .
therefore , from these data , significant time lags will only be detectable below @xmath96few
@xmath95 hz .
the time lags were computed ( for each epic detector ) for the three combinations of energy bands .
the results are shown in fig .
[ fig : phase1 ] . in each test
( three combinations of energy bands and three different detectors ) there was a significant lag at the lowest frequencies , with the softer band leading the harder band . at higher frequencies the lag becomes consistent with zero .
the lag is small , only about @xmath97 per cent between soft and hard bands ( e.g. on a timescale of 20,000 s the lag is only @xmath98 s ) .
this is illustrated in fig .
[ fig : phase1 ] , which shows the function @xmath99 compared to the measured time lags between the soft and hard bands .
the time lags were calculated using light curves for the three different detectors in identical energy ranges , to test the possible influence of poisson noise ( see section [ sect : coherence_results ] ) .
these are shown in fig .
[ fig : phase2 ] . in these tests
the lag was consistent with zero over most of the useful frequency range .
this , combined with analysis of section [ sect : phase_noise ] and the fact that the lags were recovered independently from each epic camera , suggests that the lags observed between different energy bands are not an artifact of poisson noise ( which in any event would destroy any time lag signal ) . the simulated data used in section [ sect : coh_sim ] to assess the accuracy of the coherence function were also used to assess the accuracy of the time delay spectrum .
in the first set of simulations the data in the two bands were generated using the same psd and the same random number sequence .
in addition to having unity coherence , the pairs of simulated data also have identical fourier phases and therefore zero intrinsic phase lag between them .
the top panel of fig .
[ fig : phase_sim ] shows the average of the time delay estimates , their standard deviation and the average error estimate from the 100 pairs of simulated data ( including poisson noise as described in section [ sect : coh_sim ] ) .
this average spectrum is close to the zero - delay expectation meaning that the poisson noise did not bias the time delay measurements , and the average estimated error is close to the standard deviation of the estimates . in the second set of simulations the pairs of simulated data
were generated using different psds for the two bands .
the bottom panel of fig .
[ fig : phase_sim ] shows the average time delay spectrum , which again is as expected if the delay estimate is unbiased .
it therefore seems reasonable to assume that the time delays observed in at low frequencies are intrinsic to the source .
this paper presents an analysis of the x - ray continuum variability of . a strong , linear correlation is found between the rms variability amplitude and the source flux ( section [ sect : rms - flux ] ) .
the psd was examined by fitting model power spectra to the data after allowing for the distorting effects of sampling using a monte carlo procedure ( section [ sect : pds ] ) .
however , the excellent sampling of the light curves means that these biases are in any case minimised and so there was a minimum of information lost from the psd .
the psd is well represented by a power - law of slope @xmath100 at low frequencies , breaking to a slope of @xmath101 at a frequency @xmath102 hz . however , it should be re - iterated that these psd results are model - dependent , and that parameterising the psd using a different model yields slightly different results .
the psd is energy dependent , with the variations in the higher energy band showing a flatter psd slope ( above the break frequency ) than variations in the softer bands .
there is no evidence for any energy dependence of the break frequency .
[ fig : unfolded ] illustrates the psd shape and energy dependence .
fitting the psd as a function of energy revealed tentative hints of energy - dependent , high - frequency structure in the psd ( section [ sect : e - psd ] ) .
this could be due to subtle instrument effects or may represent low - amplitude structure intrinsic to the source psd ( see also nandra & papadakis 2001 ) .
these timescales ( @xmath103 hz ) are comparable to the dynamical timescale at innermost stable orbit around the black hole . however , further progress on the very high frequency variability of relatively faint sources such as seyfert galaxies must await future high - throughput missions such as and .
the variations observed in different energy sub - bands are highly coherent at low frequencies ( long timescales ) but the coherence falls significantly below unity above @xmath96few @xmath95 hz , close to the break in the psd ( section [ sect : coherence ] ) . the high coherence between soft and hard bands at low frequencies means that the two bands are linked by a single , linear transfer function .
the direction and magnitude of the time lag shows that the hard photons lag the soft photons , on average , by @xmath104 s at a frequency @xmath105 hz ( section [ sect : phase ] ) .
it is not possible to reliably constrain the frequency dependence of the time lags , but the data seem consistent with an approximately @xmath106 dependence of the lag ( i.e. constant phase difference @xmath107 ) similar to that seen in the seyfert galaxy ngc 7469 ( papadakis 2001 ) and in gbhcs ( nowak 1999a ; pottschmidt 2000 ) . on shorter timescales
the coherence falls well below unity implying there is no simple transfer function relating soft and hard bands .
uttley ( 2002 ) presented an analysis of long timescale monitoring of using .
the psd was found to be best described by a broken power - law with slope of @xmath108 at low frequencies , breaking to @xmath109 at a frequency @xmath110 hz . using the same psd model , the break frequency found from the high - frequency data is @xmath111 hz , consistent with the result .
the slope of the psd above the break is different , but this is most likely an effect of the energy dependence of the psd . indeed , using the 210 kev data ( i.e. the same energy range used by uttley 2002 ) the slope of the high frequency psd is @xmath112 , which is consistent with the result .
the break frequency presented here ( and in uttley 2002 ) is model dependent .
this is because a model psd had to be assumed and fitted to the data to derive the appropriate parameters .
however , the broken power - law model ( with a break to @xmath108 at low frequencies ) provides consistent results when applied to the long timescale monitoring and the short timescale observation .
this , and the similarity of the observed psd to those of gbhcs ( see section [ sect : comparison ] ) suggests that the chosen model ( and therefore the derived break frequency ) is reasonable .
nowak & chiang ( 2000 ) used and data to measure the psd of .
they found the psd to break from @xmath21 to @xmath1 at a frequency @xmath14 hz and then break again to @xmath18 in the range @xmath113 hz , the position of the high frequency break , as well as the psd slopes , are roughly consistent with those found from the data .
however , as noted by uttley ( 2002 ) , the method employed by nowak & chiang ( 2000 ) did not account for distorting effects of sampling ( see section [ sect : bias ] ) and underestimated the errors at low frequencies . once these effects are included
the low - frequency flattening found by is no longer significant ( section 6.2 of uttley 2002 ) .
nowak & chiang ( 2000 ) also place a limit on any time delay between soft ( 0.52.0 kev ) and hard ( 815 kev ) bands of @xmath114 s , consistent with the time lag results presented here . hayashida ( 1998 )
also estimated the psd of , this time using data .
they found a broken power - law to be a good fit to the data .
however , these authors also did not account for the biases in the periodogram , which will be significant due to the patchy sampling of the light curves .
the unusually flat low - frequency slope ( @xmath115 ) found by these authors is most likely a result of bias in the low - frequency periodogram , and therefore the position of their claimed break frequency should be considered with some caution .
the only other agn for which detailed cross spectral analysis has been performed is ngc 7469 ( nandra & papadakis 2001 ; papadakis 2001 ) . in this object
the psd does not show any significant breaks , but does show the same energy dependence as found in .
the present analysis shows that the energy dependence of the psd is due to the slope of the psd becoming flatter at higher energies , rather than a change in the break frequency .
in addition , ngc 7469 shows time lags between energy bands in the same direction ( soft leading hard ) and with the similar relative magnitude ( @xmath97 per cent ) as in .
papadakis & lawrence ( 1995 ) showed that in ngc 4051 the harder band showed a flatter psd than the softer band . recently ,
m@xmath3hardy ( in prep . )
used data of ngc 4051 and also found this energy dependence of the pds . since first showed the medium timescale psds of seyfert galaxies to have a power - law form , their timing properties
have been compared to those of gbhcs .
now , with long timescale monitoring available from ( edelson & nandra 1999 ; nandra & papadakis 2001 ; uttley 2002 ) and short timescale observations possible with ( this paper ) it is possible to re - assess this comparison .
the best studied gbhc , cyg x-1 , when in its low / hard state shows a psd with an approximately broken power - law form .
the psd breaks from a low frequency slope of @xmath21 to @xmath1 and then to @xmath18 at high frequencies .
the position of the high frequency break is typically @xmath116 hz ( belloni & hasinger 1990 ; nowak 1999a ) . the comparable break in the psd of occurs at @xmath102 hz .
this suggests the characteristic timescales in these two systems differ by a factor @xmath117 . under the assumption that the timescales scale linearly with mass of the central black hole , and assuming cyg x-1 to contain a @xmath118 black hole ( herrero 1995 )
, then suggests the mass of the black hole in is @xmath119 ( in the range 1.55.0 @xmath120 using the 90 per cent limits on the break frequency in ) . assuming a bolometric luminosity of @xmath121 erg s@xmath8 for ( reynolds 1997 ; converting to @xmath122 km s@xmath8 mpc@xmath8 ) then implies that is radiating at the eddington luminosity .
however , since independent mass estimates ( such as from reverberation mapping ) are not available for it remains possible that this mass estimate is inaccurate if the relevant timescales do not scale linearly with black hole mass ( as assumed above ) .
if the frequency of the psd break is related to the orbital frequency of matter close to where the peak emissivity of the accretion disc occurs , @xmath123 , then it will depend on @xmath124 , where @xmath125 is the inner radius of the disc . for a standard disc , @xmath126 .
this will also introduce a dependence on @xmath127 ( where @xmath128 is the gravitational radius of the black hole ) and thus the spin parameter of the black hole @xmath129 .
if the black hole in has a higher spin parameter than that in cyg x-1 then the mass deduced from the above scaling could be increased .
the other temporal characteristics seen in are remarkably similar to those observed in cyg x-1 , implying there is indeed a strong connection between these two types of source .
the psd ( above the break frequency ) is flatter in harder energy bands in both and cyg x-1 ( nowak 1999a ; lin 2000 ) .
the strong rms - flux correlation first seen in x - ray binaries ( uttley & m@xmath130hardy 2001 ) is exhibited by ( see section [ sect : rms - flux ] ) . also displays a loss of coherence between hard and soft bands at high temporal frequencies ( i.e. above the break in the psd ) very similar to that seen in cyg x-1 ( nowak 1999a ) and time lags between these bands that appear similar ( in direction and relative magnitude ) to those observed in cyg x-1 .
in particular , the time lags measured in between soft and hard bands is @xmath98 s in the lowest frequency bin , where the coherence is near unity . in cyg
x-1 the shortest time lags observed are @xmath131 s ( nowak 1999a ; pottschmidt 2000 ) at frequencies just below where the coherence falls significantly below unity .
the ratio of these two time lags is @xmath5 , similar to the ratio between the psd break timescales .
it should also be noted that it is not yet clear whether seyfert galaxies more closely resemble the low / hard or high / soft states of gbhcs ( see also the discussion in uttley 2002 ) . in the high / soft state
the psd of cyg x-1 is steep at high frequencies ( @xmath132 ) and flattens to @xmath1 at @xmath133 hz , with no further flattening evident at lower frequencies ( cui 1997 ; churazov & gilfanov & revnivtsev 2001 ; reig , papadakis & kylafis 2002 ) .
this ratio of characteristic timescales in this case is @xmath134 and the corresponding mass estimate for is @xmath135 , implying a luminosity @xmath136 per cent of the eddington limit .
this high ( but nevertheless sub - eddington ) accretion rate might be expected if is indeed in a state similar to the high / soft state seen in cyg x-1 .
it is interesting to note that the best - fitting linear model to the rms - flux correlation observed in is consistent with passing through the zero - flux origin ( section [ sect : rms - flux ] ) .
the simplest explanation of this is if there is no significant constant component to the light curve .
a strong constant component ( @xmath137 per cent of the source flux ) is observed from the rms - flux correlation in cyg x-1 ( uttley & m@xmath3hardy 2001 ) , but only when the source is in the low / hard state ( gleissner 2002 ) .
this is perhaps also indicating that the variability properties of are more similar to the high / soft state than the low / hard state of cyg x-1 .
( the time delay properties of cyg x-1 when in its high / soft state are rather similar to those seen in its low / hard state , see e.g. pottschmidt 2000 ) .
the x - ray emission mechanism operating in seyfert galaxies is usually thought to be inverse - compton scattering of soft photons in a hot corona ( e.g. sunyaev & titarchuk 1980 ; haardt & maraschi 1991 ) . in the simplest models the harder photons are expected to lag behind the softer photons due to the larger number of scatterings required to produce harder photons , and the delay should be of order the light - crossing time of the corona .
the longest observed time delay in between soft and hard bands is @xmath98 s. assuming a black hole mass of @xmath138 , this lag corresponds to a light - crossing distance of @xmath139 ( @xmath140 ) .
thus the direction and magnitude of the observed lag in are consistent with an origin in a comptonising corona . however , if the lags are frequency dependent ( as expected by analogy with cyg x-1 and also seen in ngc 7469 ; papadakis 2001 ) the lags at lower temporal frequencies would become much longer than expected for a compact corona ( see discussion in nowak 1999b ) .
indeed , the simplest such models predict the time delay between soft and hard photons to be independent of fourier frequency ( miyamoto 1988 ) , contrary to what is observed .
in addition , some models of compton scattering coronae predict the high frequency psd should be steeper for higher energy photons , due to the high - frequency fluctuations being washed out by multiple scatterings ( hua & titarchuk 1996 ; nowak & vaughan 1996 ) , again contrary to the observations . alternatively
, the time delay between soft and hard bands could be due to the spectral evolution of individual x - ray events .
if the spectrum of the event becomes harder as the event unfolds this can lead to an average lag between soft and hard bands .
such models have been discussed by bttcher & liang ( 1998 ) and poutanen & fabian ( 1999 ) .
another alternative is that the delay between soft and hard bands could be due to a propagation effect , i.e. a `` trigger '' signal reaching the soft emission region before the hard emission region . in this scenario
the magnitude of the lag is related to the propagation speed and the distance between soft and hard emission regions .
the following section describes a model based on this kind of signal propagation . at higher temporal frequencies the coherence falls well below unity , meaning
there is no longer a simple transfer function between the soft and hard band emission . as a static corona
should maintain a transfer function ( whatever the physical mechanism ) , on timescales longer than the break in the psd ( @xmath141 s ) , where the coherence is high , the corona may be effectively static ( i.e. it maintains its average properties on these timescales ) .
on shorter timescales the corona may be dynamic and so the steep psd above the break and the loss of coherence may be due to changes in the corona itself .
the timescale of the break in the psd from @xmath142 to a steeper slope is @xmath143 s , which corresponds to a light - crossing distance of @xmath144 .
the orbital timescale at @xmath145 around a @xmath138 black hole is @xmath146 s , while the thermal timescale for a standard , thin accretion disc is @xmath147 s ( assuming a viscosity parameter @xmath148 ) , which is of the same order as the timescale of the break in the psd . the viscous timescale in such a disc is many orders of magnitude larger than these other timescales ( see the discussion of accretion disc timescales in section 5.8 of frank , king & raine 1985 ) .
however , if that part of the accretion flow responsible for modulating the x - ray emission is geometrically thick , the break in the psd could correspond to a viscous timescale ( which depends on the ratio of disc height @xmath149 to radius @xmath150 as @xmath151 ) .
various models have been considered to explain the variability of gbhcs and seyfert galaxies .
these include : various examples of `` shot noise '' ( e.g. terrell 1972 ; lehto 1989 ; merloni & fabian 2001 ) ; rotating `` hot spots '' on the surface of an accretion disc ( abramowicz 1991 ; abramowicz 1992 ; bao & stgaard 1995 ) ; occultation by moving clouds ( abrassart & czerny 2000 ) ; self - organised criticality in an accretion disc ( mineshige , ouchi & nishimori 1994 ) ; fluctuations in the accretion rate propagating through the disc ( lyubarskii 1997 ; churazov 2001 ; kotov , churazov & gilfanov 2001 ) and magnetohydrodynamic instabilities in the inner accretion disc ( hawley & krolik 2001 ) .
most of these models were developed to explain only one aspect of the variations ( e.g. its red noise nature ) .
now , however , it is possible to compare the predictions of these models with the well - determined timing properties of .
a viable physical model has to be able to reproduce not only the correct psd shape and timescales but also the other temporal characteristics such as the rms - flux correlation , the energy dependence of the psd , the time lags and the coherence .
it is not clear whether any of the above mechanism can explain all the observations , and indeed some models ( e.g. the obscuration model of abrassart & czerny 2000 ) seem to offer little hope of explaining properties such as the coherence and time lags .
a simple phenomenological model has been developed by lyubarskii ( 1997 ) , churazov ( 2001 ) , and kotov ( 2001 ) that seems to explain ( qualitatively , at least ) the temporal properties of cyg x-1 .
as the timing properties of discussed above are very similar to those of cyg x-1 , it is worth discussing this model as a possible explanation for the behaviour of seyfert galaxies . in its high / soft state
, cyg x-1 shows a @xmath142 psd over several decades in frequency ( cui 1997 ; churazov 2001 ; reig 2002 ) , down to timescales far longer than would normally be associated with the x - ray emitting region itself ( which is assumed to be compact ) .
the model put forward by lyubarskii ( 1997 ) to explain this very broad range in timescale has variations in accretion rate occurring over a large range of radii from the central black hole and propagating in towards the x - ray emitting region ( assumed to extend over only a few tens of gravitational radii ) .
if the viscosity parameter of the accretion flow is varying randomly and with the same amplitude at all radii , variations in mass accretion rate at much smaller radii will have an @xmath142 psd .
the accretion rate at the x - ray emitting region ( assumed to be concentrated at small radii ) is thus being modulated by variations produced over a large range in radius .
if the x - ray emission emerging from this region is proportional to the local accretion rate then the x - ray variations will also have an @xmath142 psd extending down to low frequencies .
each radius produces variations at a characteristic timescale comparable to the timescale for inward propagation , hence long - timescale fluctuations are produced at large radii and shorter - timescale fluctuations are produced closer in .
this works because any variations on timescales shorter than the propagation timescale are damped as they move through the disc ( see lyubarskii 1997 , and churazov 2001 ) .
this naturally explains both the @xmath142 part of the psd ( i.e. below the high frequency break ; churazov 2001 ) and the rms - flux correlation ( uttley & m@xmath130hardy 2001 ) .
once these variations reach the inner regions , where accretion energy is released as x - ray emission , variations on shorter timescales are further suppressed , leading to a steepening of the psd .
thus the high frequency break in the psd from @xmath142 to a steeper slope ( as seen in ) is due to suppression of fluctuations within the x - ray emitting region , and the timescale of the break in the psd corresponds to the characteristic timescale at the outer radius of the x - ray emitting region ( churazov 2001 ) .
the emissivity as a function of radius defines a green s function for the incoming @xmath142 fluctuations . in frequency terms , the radial extent of the x - ray producing region makes it act as a `` low - pass filter '' for the incident @xmath142 fluctuations , resulting in a break in the psd , to a steeper slope , at high frequencies .
the specific model described by churazov ( 2001 ) has a relatively stable , optically thick , geometrically thin accretion disc sandwiched by an optically thin , geometrically thick corona .
the disc produces relatively slowly varying thermal emission ( which would emerge presumably in the optical / ultraviolet for a seyfert galaxy ) while it is the variations within the corona that propagate inwards to produce the x - ray variability .
the timescale of the high frequency break thus corresponds to the accretion timescale ( @xmath96the viscous timescale ) of the geometrically thick corona at the outer edge of the x - ray emitting region .
kotov ( 2001 ) extended this model by adding the simple condition that the spectrum of the x - ray emitting region is a function of radius , with the outer regions producing a softer spectrum than the inner regions .
this model can then reproduce the observed time lags between soft and hard emission .
accretion rate variations propagating inwards reach the softer emitting region ( larger radii ) before the harder emitting region ( smaller radii ) , explaining why the soft emission leads the hard .
this can explain the energy dependence of the time lags ( as delay increases with the separation between energy bands ) as well as the frequency dependence ( since different radii correspond to different characteristic frequencies as well as different energy spectra ) .
in addition , this model explains the energy dependence of the psd above the high frequency break .
the harder emission is produced from smaller radii , which can be modulated by high frequency fluctuations , hence the harder emission will have more high frequency power than the softer emission ( i.e. the hard band psd will be flatter ) , as observed ( see fig .
[ fig : unfolded ] ) .
also , as the variations on short timescales are occurring within the x - ray emitting region , i.e. are a result of the dynamic corona , the coherence between bands will fall off at high frequencies , as mentioned in section [ sect : physics ] ( see also the discussion in nandra & papadakis 2001 ) .
such phenomenological models require that variations occurring at large radii can be fed into much smaller radii and so influence the psd .
the observed iron line profile of from this observation ( fabian 2002 ) indicates that the disc extends down to less than @xmath152 , and , from the steep inferred emissivity profile , that much of the x - ray power is generated very close to the black hole .
this presents a challenge for all such models , given that the dynamical timescale of this innermost region is @xmath118@xmath153 s , whereas the timescales measured ( e.g. the psd break timescale ) are orders of magnitude longer .
if the @xmath142 psd of is similar to that in the high / soft state of cyg x-1 ( reig 2002 ) then the variability timescales will extend to many orders of magnitude longer still .
indeed , the total power in @xmath142 spectra diverges so slowly that the psd could in principle continue as @xmath142 until very low frequencies without the total power becoming large .
( for example , model 2 from section [ sect : pds_results ] could continue down to the unphysically low frequency @xmath154 hz without the integrated power exceeding @xmath155 ! )
this paper presents a detailed time series analysis of the continuum variability of .
the psd shows a break from @xmath142 to @xmath156 at a frequency @xmath157 hz . comparing the break timescale with the analogous timescale seen in cyg x-1 , and assuming that these scale linearly with black hole mass , gives a black hole mass of @xmath6 for and suggests it is accreting at a significant fraction of the eddington limit .
the slope of the psd above the break is energy dependent , with the harder band showing more high frequency power than the softer band
. the variations in different energy bands also become incoherent above the break in the psd .
this leads us to speculate that on timescales shorter than the break timescale , the x - ray emitting corona is dynamic , and that harder photons are produced at smaller radii than softer photons . at the lowest fourier frequencies the soft and hard bands are highly coherent and the soft band
is found to lead the hard band by @xmath104 s. the direction and magnitude of this lag is consistent with simple comptonisation models .
a phenomenological model , originally developed to explain the @xmath142 psd of gbhcs such as cyg x-1 , is discussed and found to be able to reproduce ( qualitatively ) the observed temporal characteristics of .
although this model is rather simple and phenomenological it can explain the observed temporal characteristics of the x - ray continuum in .
based on observations obtains with , an esa science mission with instruments and contributions directly funded by esa member states and the usa ( nasa ) .
we are grateful to phil uttley for many valuable discussions on power spectral issues and to mike nowak for discussions on coherence estimation .
sv acknowledges many valuable discussions with p. thomas and financial support from pparc .
we thank an anonymous referee for useful comments .
abramowicz m. a. , bao g. , lanza a. , zhang x .- h . , 1991 , a&a , 245 , 454 abramowicz m. a. , 1992 , in s. holt , s. g. neff , c. m. urry eds .
_ testing the agn paradigm _ , p69 abrassart a. , czerny b. , 2000 , a&a , 356 , 475 belloni t. , hasinger g. , 1990 , a&a , 227 , l33 belloni t. , psaltis d. , van der klis m. , 2002 , apj , 572 , 392 bendat j. s. , piersol a. g. , 1986 , _ random data : analysis and measurement procedures _ , wiley ( new york ) bevington p. r. , robinson d. k. , 1992 , _ data reduction and error analysis for the physical sciences _ , mcgraw - hill ( new york ) bloomfield p. , 2000 , _ fourier analysis of time series _ , wiley ( new york ) bao g. , stgaard e. , 1995 , apj , 443 , 54 bttcher m. , liang e. p. , 1998 ,
apj , 506 , 281 brandt w. n. , boller th . , fabian a. c. , ruszkowski m. , 1999 , mnras , 303 , l58 churazov e. , gilfanov m. , revnivtsev m. , 2001 , a&a , 321 , 759 cui w. , zhang s. n. , focke w. , swank j. h. , apj , 484 , 383 deeter j. e. , boynton p. e. , 1982 , apj , 261 , 337 done c. , madejski g. m. , mushotzky r. f. , turner t. j. , koyama k. , kunieda h. , 1992 , apj , 400 , 138 edelson r. , nandra k. , 1999 , apj , 514 , 682 edelson r. , turner t. j. , pounds k. a. , vaughan s. , markowitz a. , marshall h. , dobbie p. , warwick r. s. , 2002 , apj , 568 , 610 fabian a. c. , 1979 , proc .
london , ser .
a , 366 , 449 fabian a. c. 2002 , mnras , 335 , l1 frank j. , king a. , raine d. , 1985 , _ accretion power in astrophysics _ , cambridge univ . press ( cambridge ) gleissner t. , wilms j. , pottschmidt k. , uttley p. , nowak m. a. , staubert r. , 2002 , in ph .
durouchoux , y. fuchs and j. rodriguez eds .
_ proceedings of the 4th microquasar workshop _ , in press ( astro - ph/0207610 ) green a. r. , m@xmath3hardy i. m. , lehto h. j. , 1993 , mnras , 265 , 664 green a. r. , m@xmath3hardy i. m. , done c. , 1999 , mnras , 305 , 309 haardt f. , maraschi l. , 1991 , apj , 380 , l51 hawley j. f. , krolik j. h. , 2001 , apj , 548 , 348 herrero a. , kudritzki r. p. , gabler r. , vilchez j. m. , gabler a. , 1995 , a&a , 297 , 556 hayashida k. , miyamoto s. , kitamoto s. , negoro h. , inoue h. , 1998 , apj , 500 , 642 jansen f. 2001 , a&a , 365 , l1 jenkins g. m. , watts , d. g. , 1968 , _ spectral analysis and its applications _ , holden - day ( san fancisco ) kotov o. , churazov e. , gilfanov m. , 2001 , mnras , 327 , 799 lampton m. , margon b. , bowyer s. , 1976 , apj , 208 , 177 lawrence a. , papadakis i. , 1993 , apj , 414 , l85 leahy d. a. , darbro w. , elsner r. f. , weisskopf m. c. , kahn s. , sutherland p. g. , grindlay j. e. , 1983 , apj , 266 , 160 lehto h. j. , 1989 , in j. hunt , b. battrick , eds , _ two topics in x ray astronomy _ , ( esa sp-296 ; noordwijk : esa ) , p499 lin d. , smith i. a. , bttcher m. , liang e. p. , 2000 ,
apj , 531 , 963 lomb n. r. , 1976 , ap&ss , 39 , 447 lyubarskii y. e. 1997 , mnras , 292 , 679 markowitz a. 2002 , apj , submitted m@xmath3hardy i. m. , 1989 , in j. hunt , b. battrick , eds , _ two topics in x ray astronomy _ , ( esa sp-296 ; noordwijk : esa ) , p1111 merloni a. , fabian a. c. , 2001 , mnras , 328 , 958 mineshige s. , ouchi n. b. , nishimori h. , 1994 , pasj , 46 , 97 miyamoto s. , kitamoto s. , mitsua k. , dotani t. , 1998 , nature , 336 , 450 miyamoto s. , kitamoto s. , 1989 , nature , 342 , 773 miyamoto s. , kimura k. , kitamoto s. , dotani t. , ebisawa k. , 1991 , apj , 383 , 784 mushotzky r. f. , done c. , pounds k. a. , 1993 , ara&a , 31 , 717 nandra k. , papadakis i. e. , 2001 , apj , 554 , 710 nowak m. a. , vaughan b. a. , 1996 , mnras , 280 , 227 nowak m. a. , vaughan b. a. , wilms j. , dove j. b. , begelman m. c. , 1999a , apj , 510 , 874 nowak m. a. , wilms j. , vaughan b. a. , dove j. b. , begelman m. c. , 1999b , apj , 515 , 726 nowak m. a. , chiang j. , 2000 , apj , 531 , l13 nowak m. a. , wilms j. , dove j. b. , 2002 , mnras , 332 , 856 papadakis i. e. , lawrence a. , 1993 , mnras , 261 , 612 papadakis i. e. , lawrence a. , 1995 , mnras , 272 , 161 papadakis i. e. , nandra k. , kazanas d. , 2001 , apj , 554 , l133 poutanen j. , fabian a. c. , 1999 , mnras , 306 , l31 pottschmidt k. , wilms j. , nowak m. a. , heindl w. a. , smith d. m. , staubert r. , 2000 , a&a , 357 , l17 press w. h. 1978 , comments on astrophysics , 7 , 103 press w. h. , rybicki g. b. , 1989 , apj , 338 , 277 press w. h. , teukolsky s. a. , vetterling w. t. , flannery b. p. , 1992
, _ numerical recipes
_ , cambridge univ . press .
( cambridge ) priestley m. b. , 1981 , _ spectral analysis and time series _ , academic press ( london ) rees m. j. , 1984 , ara&a , 22 , 471 reig p. , papadakis i. , kylafis n. d. , 2002 , a&a , 383 , 292 reynolds c. s. , ward m. j. , fabian a. c. , celotti a. , 1997 , mnras , 291 , 403 scargle j. d. , 1982 , apj , 263 , 835 stella l. , arlandi e. , tagliaferri g. , israel g. l. , 1997 , in t. subba rao , m. b. priestley , o. lessi eds . , _
applications of time series analysis in astronomy and meteorology _ , chapman & hall ( london ) ( astro - ph/9411050 ) strder l. 2001 , a&a , 365 , l18 sunyaev r. a. , titarchuk l. g. , 1980 , a&a , 86 , 121 terrell n. j. , 1972 , apj , 174 , l35 thorne k. s. , 1974 , apj , 191 , 507 timmer j. , knig m. , 1995 , a&a , 300 , 707 uttley p. , m@xmath3hardy i. m. , 2001 , mnras , 323 , 26 uttley p. , m@xmath3hardy i. m. , papadakis i. , 2002 , mnras , 332 , 231 van der klis m. , 1989 , in h. ogelman , e. p. j. van den heuvel eds . , _ timing neutron stars _ , kluwer ( dordrecht ) , nato asi series c 262 , p27 van der klis m. , 1995 , in w. h. g. lewin , j. van paradijs , e. p. j. van den heuvel eds . , _ x - ray binaries _ , cambridge univ . press ( cambridge ) , p252 van der klis m. , 1997 , in g.j .
babu , e. d. feigelson eds . , _
statistical challenges in modern astronomy ii _ , springer - verlag ( new york ) , p321 vaughan b. a. , nowak m. a. , 1997 , apj , 474 , l43 | this paper presents a comprehensive examination of the x - ray continuum variability of the bright seyfert 1 galaxy .
the source clearly shows the strong , linear correlation between rms variability amplitude and flux first seen in galactic x - ray binaries . the high frequency power spectral density ( psd ) of
is examined in detail using a monte carlo fitting procedure and is found to be well represented by a steep power - law at high frequencies ( with a power - law index @xmath0 ) , breaking to a flatter slope ( @xmath1 ) below @xmath2 hz , consistent with the previous results of uttley , m@xmath3hardy & papadakis .
the slope of the power spectrum above the break is energy dependent , with the higher energies showing a flatter psd . at low frequencies the variations between different energy bands are highly coherent while at high frequencies the coherence is significantly reduced .
time lags are detected between energy bands , with the soft variations leading the hard .
the magnitude of the lag is small ( @xmath4 s for the frequencies observed ) and is most likely frequency dependent .
these properties are remarkably similar to the temporal properties of the galactic black hole candidate cygnus x-1 .
the characteristic timescales in these two types of source differ by @xmath5 ; assuming that these timescales scale linearly with black hole mass then suggests a black hole mass @xmath6 for .
we speculate that the timing properties of may be analogous to those of cyg x-1 in its high / soft state and discuss a simple phenomenological model , originally developed to explain the timing properties of cyg x-1 , that can explain many of the observed properties of . [ firstpage ] galaxies : active galaxies : seyfert : general galaxies : individual : x - ray : galaxies |
in the study of moduli spaces , it is a problem to provide compactifications which are functorial and with meaningful geometric and combinatorial properties .
a leading example in this sense is the deligne - mumford compactification of the moduli space of stable @xmath3-pointed curves .
another successful example is alexeev s compactification of the moduli space of principally polarized abelian varieties @xmath4 using stable semiabelic pairs ( see @xcite ) .
compactifications of the moduli space of polarized k3 surfaces also received a lot of attention ( see @xcite ) .
this motivates us to consider enriques surfaces , together with the compactification method provided by the theory of stable pairs ( see @xcite and @xcite for recent developments ) . in this paper
we focus on the @xmath5-polarized enriques surfaces , which were considered in @xcite .
@xmath5-denotes the sublattice of @xmath6 of vectors of even square .
a @xmath5-polarization on an enriques surface @xmath7 is a primitive embedding of @xmath5 into @xmath8 with some additional requirements ( see definition [ d16polarizedenriquessurfaces ] ) .
these enriques surfaces can be realized as the @xmath1-cover of the blow up of @xmath2 at three general points branched along a configuration of six lines .
our choice of divisor consists of the ramification divisor multiplied by a rational number @xmath9 , @xmath10 .
let @xmath11 be the closure in the coarse moduli space of stable surface pairs of the locus of points parametrizing these enriques surfaces with our choice of divisor ( see section [ constructionoftheksbacompactification ] ) .
[ firstmainresult ] the stratification of the boundary of @xmath11 and the degenerations parametrized by it are described in theorem [ combinatoricsstratificationstatement ] .
the boundary has two divisorial components , one of codimension @xmath12 , and three @xmath13-dimensional strata .
the compactification is toroidal , baily - borel , and a mixture of the two in a neighborhood of each @xmath13-dimensional stratum respectively .
we provide two combinatorial interpretations of the toroidal part of the boundary ( see figures [ subdivisionscubewithoutcornercuts ] , [ coxeterdiagramandsubdiagrams ] ) .
let us explain the techniques used to prove this result .
denote by @xmath14 the toric boundary of @xmath15 and let @xmath16 be a generic effective divisor of class @xmath17 .
then the @xmath1-cover of @xmath18 branched along @xmath19 is an enriques surface with two choices of @xmath5-polarization , and we can reduce ourselves to consider the pairs @xmath20 ( this reduction is explained in section [ enriquesandchoicedivisor ] ) .
the advantage of this setting is that if @xmath21 denotes the unit cube , then @xmath22 is a stable toric pair of type @xmath23 ( see definition [ definitionofstabletoricpair ] ) , and therefore we can take advantage of the moduli theory for such pairs , which is developed in @xcite .
let @xmath24 be the projective coarse moduli space parametrizing @xmath22 and its degenerations .
denote by @xmath25 the normalization of @xmath26 .
the next result summarizes theorems
[ allyouneedtoknowaboutksbapoints ] and [ propertiesofthemapstptoksba ] . [ secondmaintheorem ] if @xmath27 denotes the symmetry group of @xmath21 , then @xmath28 . on a dense open subset ,
the isomorphism maps the @xmath27-class of a stable toric pair @xmath29 to @xmath20 , where @xmath14 denotes the toric boundary of @xmath30 .
one of the main problems in the proof of theorem [ secondmaintheorem ] is to compute the image of @xmath29 when @xmath20 is unstable , which happens if and only if the polyhedral subdivision of @xmath31 associated to @xmath29 has a _ corner cut _ ( see definition [ definitionofcornercut ] ) .
this aspect is analyzed in section [ modificationsofthepairs ] , were we describe a modification of @xmath29 , which we denote by @xmath32 , such that @xmath33 is a stable pair .
moreover , this modification can be performed in a one - parameter family . at this point ,
given the isomorphism of theorem [ secondmaintheorem ] , we can explicitly describe @xmath11 because the moduli space @xmath25 is the projective toric variety associated to the secondary polytope of @xmath21 , and it is stratified according to the ( regular ) polyhedral subdivisions of @xmath21 . to conclude , we consider the baily - borel compactification @xmath34 , where @xmath35 is the period domain parametrizing @xmath5-polarized enriques surfaces ( this compactification was studied in @xcite ) , and we compare it to @xmath36 . we construct a birational morphism @xmath37 extending the period map to the boundary of @xmath36
( see theorem [ morphismtobailyborelcompactification ] ) .
the double nature , namely toroidal and baily - borel , of the boundary of the moduli space @xmath36 produced by the theory of stable pairs is remarkable : it illustrates what is the behavior one should expect in general when considering compactifications by stable pairs ( see also ( * ? ? ?
* remark 1.5 ) ) .
moduli of configurations of lines on @xmath2 and on the blow up of @xmath2 at three general points , and branched covers of these , are object of wide interest in the literature ( see @xcite ) .
a compact moduli space for del pezzo surfaces using stable pairs was constructed in @xcite , but in our case we have a different choice of the divisor .
our work is also related to @xcite in the case of sextic plane curves , but here we consider the blow up @xmath2 . some of the techniques we use in this paper are closely related to @xcite ( for instance the use of stable toric pairs and polyhedral subdivisions ) .
the constructions and the results of the current paper can be adapted to describe a compactification by stable pairs of the whole @xmath38-dimensional family of enriques surfaces .
this will be object of future work . in section [ stabletoricpairsstablepairs ]
we recall the needed background about stable toric pairs and stable pairs .
this theory is immediately applied in section [ enriquesandchoicedivisor ] , where we define the @xmath5-polarized enriques surfaces and the surface pairs we want to study . here
we also define our moduli spaces of interests : @xmath24 and @xmath11 .
section [ modificationsofthepairs ] is devoted to the definition of @xmath32 for a stable toric pair @xmath29 , and then in section [ wallcrossingstudy ] we prove that @xmath33 is stable for every stable toric pair @xmath29 . in section [ sectiononksbacompactification ]
we first construct a morphism @xmath39 , and then we prove the isomorphism @xmath40 , which allows us to describe the stable pairs parametrized by the boundary of @xmath11 and its stratification .
finally , in section [ sectiononksbacompactification ] we relate @xmath11 with the baily - borel compactification @xmath34 .
we work over an arbitrary algebraically closed field @xmath41 of characteristic zero , exception made for section [ relationwithbailyborel ] where we work over @xmath42 .
i want to express my gratitude to my advisor , valery alexeev , for guiding me through this project and for his help .
i am also grateful to robert varley for many discussions related to this work .
many thanks to patricio gallardo for great suggestions and to adrian brunyate for helpful conversations in an earlier stage of this project .
i gratefully acknowledge financial support from the office of the graduate school of the university of georgia , the nsf grant dms-1603604 , and the research and training group in algebra , algebraic geometry , and number theory , at the university of georgia .
in this section we review some basic definitions and results in the theory of moduli of stable toric pairs and stable pairs .
our main references are @xcite . for the terminology we use from convex geometry ( especially for the definition of the secondary polytope )
we refer to @xcite . for us ,
a _ variety _ is a connected and reduced scheme of finite type over an extension @xmath43 of @xmath41 , which is our fixed base field of characteristic zero ( in particular , a variety need not be irreducible ) .
@xmath44 denotes a fixed ( split ) torus over @xmath43 and let @xmath45 be its character lattice .
@xmath46 denotes the tensor product @xmath47 .
we want to remark that the theory of stable toric pairs we are about to review works in any characteristic .
let @xmath30 be a variety with a @xmath44-action .
we say that @xmath30 is a _ stable toric variety
_ if @xmath30 is seminormal and its irreducible components are toric varieties under the @xmath44-action .
the _ toric boundary _ of a stable toric variety is defined to be the sum of the boundary divisors of each irreducible component which are not in common with other irreducible components .
if @xmath30 is projective and @xmath48 is an ample and @xmath44-linearized invertible sheaf on @xmath30 , we say that the pair @xmath49 is a _ polarized _ stable toric variety .
[ combinatorialdescriptionpolarizedstabletoricvariety ] assume we have a polarized stable toric variety @xmath49 .
then we can associate to each irreducible component @xmath50 of @xmath30 a lattice polytope @xmath51 .
these polytopes can be glued to one another in the same way as @xmath30 is the union of its irreducible components .
this results into a topological space @xmath52 which is called the _ topological type _ of @xmath30 .
the topological type comes together with a finite map @xmath53 , called the _ reference map _ , which embeds each @xmath51 as a lattice polytope in @xmath46 .
the set of the faces of all the polytopes @xmath51 , together with the identifications coming from the gluing , forms what is called a _
complex of polytopes_. up to isomorphism , there is a @xmath54-to-@xmath54 correspondence between polarized stable toric varieties ( for a fixed torus @xmath44 ) and the following data : * a complex of polytopes @xmath55 ; * a reference map @xmath56 ; * an element of the cohomology group we are about to describe . for each @xmath57 ,
let @xmath58 be the saturated sublattice of @xmath59 generated by @xmath60 and let @xmath61 be the torus @xmath62 .
the tori @xmath61 for @xmath57 define a sheaf @xmath63 on the poset @xmath55 with the order topology .
so we want to consider an element of @xmath64 , which describes the way the irreducible components of @xmath30 are glued together . for more details about this , see ( * ? ? ?
* section 4.3 ) or ( * ? ? ?
* theorem 1.2.6 ) .
let @xmath49 be a polarized stable toric variety and let @xmath65 be a lattice polytope .
we say that @xmath30 has _ type @xmath23 _ if the complex of marked polytopes @xmath55 associated to @xmath30 is a polyhedral subdivision of the marked polytope @xmath66 .
observe that in this case the reference map is implicitly given by the inclusion of @xmath21 in @xmath46 .
furthermore , the toric boundary of @xmath30 is given by the sum of the divisors corresponding to the facets in @xmath55 contained in the boundary of @xmath21 .
[ definitionofstabletoricpair ] a _ stable toric pair _ is the datum of a polarized stable toric variety @xmath49 together with an effective cartier divisor @xmath18 which is the divisor of zeros of a global section of @xmath48 .
also , we require that @xmath18 does not contain any torus fixed point ( or equivalently any @xmath44-orbit ) .
we denote a stable toric pair simply by @xmath29 because it is understood that @xmath67 .
two stable toric pairs @xmath29 and @xmath68 are isomorphic if there exists an isomorphism @xmath69 that preserves the @xmath44-action and such that @xmath70 .
we say that a stable toric pair has type @xmath23 if the corresponding polarized stable toric variety has type @xmath23 .
[ combinatorialdescriptionstabletoricpair ] let @xmath29 be a stable toric pair and let @xmath55 be the complex of polytopes corresponding to the polarized stable toric variety @xmath71 .
if @xmath50 is any irreducible component of @xmath30 , then the restriction @xmath72 can be described combinatorially as follows ( see ( * ? ? ?
* theorem 1.2.7 ) in combination with ( * ? ? ?
* lemma 2.2.7 , part 2 ) ) .
consider the marking on the corresponding lattice polytope @xmath51 given by @xmath73 .
observe that these lattice points in @xmath51 correspond to monomials .
now , @xmath72 is determined ( not uniquely ) by a function @xmath74 , which assigns to each monomial a corresponding coefficient ( which we want to be nonzero because @xmath18 does not contain any torus fixed point ) .
by @xmath75 we denote the category of locally noetherian schemes over our fixed base field @xmath41 . given @xmath76 , a _ family of stable toric pairs _
@xmath77 is the datum of a proper and flat morphism of schemes @xmath78 , a compatible @xmath79 action on @xmath80 and an effective cartier divisior @xmath81 such that @xmath82 is flat and the fiber @xmath83 over every geometric point @xmath84 is a stable toric pair with the action induced by @xmath85 ( this makes sense because in section [ stabletoricpairs ] we defined stable toric pairs over an arbitrary field extension of @xmath41 ) .
two families of stable toric pairs over the same base are _ isomorphic _ if there exists an isomorphism of pairs over @xmath7 preserving the torus action . given a lattice polytope @xmath21 , we say that a family of stable toric pairs has _
type @xmath23 _ if every geometric fiber has type @xmath23 .
[ algebraicdatafamilyofstabletoricpairs ] let @xmath77 be a family of stable toric pairs .
denote by @xmath86 the morphism @xmath78 and let @xmath87 . following ( * ? ? ?
* proof of lemma 2.10.1 ) , for @xmath88 the sheaves @xmath89 are locally free , and the torus action gives a decomposition @xmath90 into sheaves that are also locally free .
this results into a locally free @xmath91-graded @xmath92-algebra @xmath93 together with a @xmath94-degree @xmath54 section @xmath95 of @xmath96 corresponding to @xmath97 .
conversely , a pair @xmath98 where @xmath96 is a locally free graded @xmath92-algebra and @xmath95 a degree @xmath54 section uniquely determines a family of stable toric pairs up to isomorphism ( see ( * ? ? ?
* proof of theorem 2.10.8 ) or ( * ? ? ?
* section 4 ) ) . after defining families of stable toric pairs
we automatically have a notion of stack over @xmath41 . given a lattice polytope @xmath21 ,
denote by @xmath99 the category of families of stable toric pairs of type @xmath23 , where a morphism @xmath100 is a pullback diagram ( a)[matrix of math nodes , row sep=2em , column sep=1.5em , text height=1.5ex , text depth=0.25ex ] ( , ) & ( , ) + s&s .
+ ; ( a-1 - 1 ) edge node[above]@xmath101(a-1 - 2 ) ; ( a-1 - 1 ) edge node[left](a-2 - 1 ) ; ( a-2 - 1 ) edge node[below]@xmath102(a-2 - 2 ) ; ( a-1 - 2 ) edge node[right](a-2 - 2 ) ; let @xmath103 be the functor sending @xmath77 to @xmath7 and a morphism @xmath104 to @xmath102 .
then @xmath99 is a stack over @xmath41 parametrizing families of stable toric pairs of type @xmath23 .
the main result about about the stack @xmath99 is the following theorem due to alexeev .
[ modulistabletoricpairs ] let @xmath21 be a lattice polytope and let @xmath99 be the stack of stable toric pairs of type @xmath23 .
then the following hold : * @xmath99 is a proper quotient stack ( and therefore an artin stack ) with finite stabilizers , we have that @xmath99 is actually a deligne - mumford stack . ] ; * it has a coarse moduli space @xmath24 which is a projective scheme ; * @xmath105 is naturally stratified , and every stratum corresponds in a @xmath54-to-@xmath54 way to a polyhedral subdivision of @xmath66 ; * the normalization of the main irreducible component of @xmath106 is isomorphic to the toric variety associated to the secondary polytope @xmath107 .
let @xmath30 be a variety and let @xmath108 be a divisor on @xmath30 where , for all @xmath109 , @xmath110\cap\mathbb{q}$ ] and @xmath111 is a prime divisor .
then the pair @xmath29 is _ semi - log canonical _ if the following conditions are satisfied : * @xmath30 is @xmath112 and every codimension @xmath54 point is regular or a double crossing singularity ( varieties satisfying these properties are also called _ demi - normal _ ) ; * if @xmath113 is the normalization with conductors @xmath114 and @xmath115 ( see ( * ? ? ?
* section 5.1 ) ) , then the support of @xmath18 does not contain any irreducible component of @xmath116 ; * @xmath117 is @xmath118-cartier ; * the pair @xmath119 is _ log canonical _ ( see ( * ? ? ? * definition 2.34 ) ) . a pair @xmath29 is _ stable _ if the following conditions are satisfied : * _ on singularities : _ @xmath29 is a semi - log canonical pair ; * _ numerical : _
@xmath117 is ample .
the viehweg s moduli stack @xmath120 is defined as follows .
let us fix constants @xmath121 , @xmath122 and @xmath123 with @xmath110\cap\mathbb{q}$ ] and @xmath124 for all @xmath125 . for any scheme @xmath7 over @xmath41 , @xmath126 is the set of proper flat families @xmath78 together with a divisor @xmath127 satisfying the following properties : * for all @xmath125 , @xmath128 is a codimension @xmath54 closed subscheme which is flat over @xmath7 ; * every geometric fiber @xmath29 is a stable pair of dimension @xmath129 with @xmath130 ; * there exists an invertible sheaf on @xmath48 on @xmath80 such that for every geometric fiber @xmath29 one has @xmath131 . whether or not the stack @xmath120 is coarsely represented by a projective scheme is a subtle matter . in our case of interest
( see section [ constructionoftheksbacompactification ] ) it is .
more details can be found in @xcite .
for us a surface is a @xmath132-dimensional projective variety . recall that an _ enriques surface _
@xmath133 is a smooth irreducible surface with @xmath134 and @xmath135 ( these assumptions are enough to imply that @xmath7 is minimal with kodaira dimension @xmath13 and @xmath136 ) .
let @xmath137 denote the standard lorentzian lattice of signature @xmath138 .
then the @xmath5 lattice is the index @xmath132 sublattice of @xmath137 containing vectors with even square ( or equivalently , we can require the sum of the coordinates to be even ) .
if @xmath139 is the canonical basis of @xmath137 , then we distinguish the following vectors in @xmath5 : @xmath140 and @xmath141 .
a @xmath5-polarized enriques surface is an enriques surface whose picard group contains a primitively embedded copy of @xmath5 such that : * the distinguished vector @xmath142 corresponds to a nef divisor class @xmath143 ( observe that @xmath143 is also big because @xmath144 ) ; * the distinguished vectors @xmath145 and @xmath141 correspond to six irreducible curves @xmath146 and @xmath147 respectively ( these curves are @xmath148-curves , i.e. isomorphic to @xmath149 and with self - intersection @xmath150 ) .
the next proposition is well known from ( * ? ? ?
* section 3 ) .
however , given its importance for us , we briefly sketch its proof .
let @xmath7 be a @xmath5-polarized enriques surface .
then the linear system @xmath151 is a genus @xmath54 pencil , @xmath152 .
the linear system @xmath153 is base point free and of dimension @xmath132 ( see ( * ? ? ?
* proposition 3.1 ) ) .
moreover , @xmath153 contracts the distinguished curves @xmath154 for @xmath152 because @xmath155 . on the other hand
, @xmath151 is a proper linear subsystem of @xmath153 ( for instance , it does not contract @xmath156 and contracts @xmath157 for @xmath158 ) .
therefore @xmath151 is a pencil and a curve @xmath159 in this linear system has ( arithmetic ) genus @xmath54 because @xmath160 ( now use ( * ? ? ?
* i.15 the genus formula ) ) .
[ hereisthepairofinterest ] every genus @xmath54 fibration on an enriques surface has exactly two double fibers which are called _ half - fibers _ or _ half - pencils _ ( see ( * ? ? ? * chapter viii , section 17 ) ) .
therefore , for a @xmath5-polarized enriques surface @xmath7 , denote by @xmath161 and @xmath162 the two half - fibers in the genus @xmath54 pencil @xmath151 , @xmath152 .
then the divisor @xmath163 is canonically associated to the @xmath5-polarization , and its intrinsic nature leads us to consider moduli of pairs @xmath164 for @xmath10 , @xmath165 . by ( * ? ? ?
* theorem 3 ) we know that every enriques surface contains three genus @xmath54 pencils @xmath166 such that @xmath167 . [ blowupatthreepointsanddivisor ]
let @xmath168 be the blow up of @xmath2 at @xmath169,[0:1:0]$ ] and @xmath170 $ ] .
@xmath168 comes with three genus @xmath13 pencils @xmath171 and denote by @xmath172 two distinct irreducible elements in the @xmath109-th pencil , @xmath152 .
[ d16asz22coverandconverse ] let @xmath7 be a @xmath5-polarized enriques surface .
then there exists a divisor @xmath173 on @xmath174 ( see notation [ blowupatthreepointsanddivisor ] ) without triple intersection points and a morphism @xmath175 such that * @xmath86 is the @xmath1-cover of @xmath174 ramified at @xmath163 and branched along @xmath173 ; * the linear system @xmath176 is the pullback of @xmath177 , @xmath152 .
conversely , given any divisor @xmath173 on @xmath174 without triple intersection points , the @xmath1-cover of @xmath174 branched along @xmath173 is an enriques surface with exactly two choices of @xmath5-polarization : @xmath143 is the pullback of a generic line in @xmath174 and the six lines @xmath156 , @xmath152 , are the preimages of the exceptional divisors in @xmath174 , or the other three @xmath178-curves .
these two choices are exchanged by the cremona involution of @xmath174 .
\(i ) and ( ii ) are discussed in ( * ? ? ?
* section 3 ) .
the divisor @xmath173 can not have triple intersection points otherwise @xmath7 would be singular ( see ( * ? ? ?
* proposition 3.1 ) ) . for the inverse construction
, we use the general theory of abelian covers developed in @xcite .
let @xmath179 where @xmath180 is the identity element , and let @xmath181 be the characters of @xmath1 with @xmath182 and @xmath183 . define @xmath184 and @xmath185 .
consider the building data ( see ( * ? ? ?
* definition 2.1 ) ) consisting of the divisors @xmath186 and the line bundles @xmath187 satisfying @xmath188 this building data determines a @xmath1-cover @xmath189 branched along @xmath173 which is unique up to isomorphism of @xmath1-covers ( see ( * ? ? ? * theorem 2.1 ) ) . by (
* proposition 3.1 ) we have that @xmath7 is smooth , and using ( * ? ? ?
* proposition 4.2 , formula ( 4.8 ) ) one can easily compute that @xmath190 , which tells us that @xmath191 . if @xmath96 denotes the ramification divisor of the cover ,
then @xmath192 and @xmath193 imply that @xmath194 , hence @xmath7 is an enriques surface .
[ reductiontobaseofthecover ] let @xmath7 be a @xmath5-polarized enriques surface and let @xmath175 be the corresponding @xmath1-cover ramified at @xmath195 and branched along @xmath196 ( see proposition [ d16asz22coverandconverse ] ) .
then @xmath197 in particular , @xmath198 is stable if and only if @xmath199 is stable .
we have that @xmath200 and @xmath201 .
this implies that @xmath202 , and by adding @xmath203 to both sides we obtain what claimed . for the last statement about stability
, we have that @xmath198 is semi - log canonical if and only if @xmath204 is semi - log canonical by ( * ? ? ? * lemma 2.3 ) , and @xmath205 is ample if and only if @xmath206 is ample by ( * ? ? ?
* proposition 1.2.13 and corollary 1.2.28 ) .
[ equivalenttoconsiderthebaseofthecover ] with the notation introduced in lemma [ reductiontobaseofthecover ] , we claim that studying the degenerations of the stable pairs @xmath198 is equivalent to the study of the degenerations of the stable pairs @xmath204 .
to prove this , let @xmath207 be the field of fractions of a dvr @xmath208 , where @xmath209 is the maximal ideal of @xmath210 .
let @xmath211 ( resp .
@xmath212 ) be a family of stable pairs over @xmath213 with fibers isomorphic to @xmath198 ( resp .
@xmath204 ) . let @xmath214 be the @xmath1-cover ramified at @xmath215 and branched along @xmath216 .
let @xmath217 be the completion of @xmath211 over @xmath218 , or a ramified base change of it , having as central fiber a stable pair ( see ( * ? ? ?
* theorem 2.1 ) ) .
denote by @xmath219 the central fiber of @xmath217 .
we adopt similar notations for @xmath212 . what we need to show
is that @xmath219 is the @xmath1-cover of @xmath220 .
this is automatic if we can show that the @xmath1-action on @xmath221 extends to @xmath222 ( in this case , the quotient of @xmath222 by @xmath1 is isomorphic to @xmath97 by the uniqueness of the completion of @xmath223 over @xmath218 , and this implies what we want ) . fix any @xmath224 and consider its corresponding action @xmath225 . resolving the indeterminacies of @xmath226
, we obtain a morphism @xmath227 where @xmath228 is obtained by blowing up @xmath222
. then @xmath229 corresponds to a morphism @xmath230 from the log canonical model of @xmath228 to the log canonical model of @xmath222 , which are both isomorphic to @xmath222 .
then @xmath231 is the desired extension of @xmath232 .
let @xmath233,[y_0:y_1],[z_0:z_1])$ ] be coordinates in @xmath15 .
we denote by @xmath14 the toric boundary @xmath234 ( @xmath235 denotes the closed subscheme associated to the homogeneous ideal @xmath236 ) . [ mainconstruction ] consider @xmath174 together with a divisor @xmath173 ( see notation [ blowupatthreepointsanddivisor ] ) without triple intersection points .
then there exists @xmath237 with coefficients @xmath238 such that @xmath239 is isomorphic to @xmath240 .
moreover , such @xmath16 is uniquely determined up to the action of @xmath241 , where @xmath21 is the unit cube .
consider the three projections @xmath242 , @xmath152 .
let @xmath243)$ ] and @xmath244)$ ] .
the morphism @xmath245 is an embedding whose image is a divisor of class @xmath17 ( observe that the restriction of @xmath14 to this divisor gives the six @xmath178-curves , each one with multiplicity @xmath132 ) . for each one of the three copies of @xmath149 ( which we label with @xmath152 ) choose an automorphism @xmath246 sending @xmath247,[a_{0i}':a_{1i}']$ ] to @xmath248,[0:1]$ ] respectively
let @xmath18 be the image of the composition of the embedding @xmath249 followed by the automorphism @xmath250 . then , under this morphism , @xmath251 and @xmath173 corresponds to @xmath19 .
moreover , @xmath252 where the coefficients @xmath253 are nonzero ( otherwise @xmath19 would have triple intersection points ) . in the construction of @xmath18 above we made some choices .
there is a @xmath254-action which comes from the fact that we can permute the three projections @xmath255 and for each @xmath109 we can exchange @xmath247 $ ] with @xmath256 $ ] ( @xmath254 is isomorphic to the wreath product @xmath257 ) .
there is also a @xmath258-action due to the fact that each @xmath246 is uniquely determined up to @xmath259 .
it is easy to observe that our construction of @xmath18 is invariant under the action of @xmath260 on the line arrangement ( see ( * ? ? ?
* theorem 8.4.2 ) for the description of @xmath260 ) . to conclude we need to show that any realization @xmath237 with coefficients @xmath238 such that @xmath261 can be obtained as we described up to @xmath262-action .
but this is true because , up to @xmath263 , there is a unique way to realize @xmath168 in @xmath15 so that the six @xmath178-curves are given by the restriction of @xmath14 , and this is given by @xmath264 ( we omit the proof of this ) .
summing up , starting from our pairs of interest @xmath265 defined in observation [ hereisthepairofinterest ] , the considerations we made so far ( proposition [ d16asz22coverandconverse ] , lemma [ reductiontobaseofthecover ] , remark [ equivalenttoconsiderthebaseofthecover ] and proposition [ mainconstruction ] ) allow us to consider moduli of the pairs @xmath20 with @xmath16 as prescribed by proposition [ mainconstruction ] ( observe that @xmath20 is obviously stable ) .
this approach in terms of @xmath16 is convenient because @xmath22 is a stable toric pair of type @xmath23 where @xmath21 is the unit cube , and theorem [ modulistabletoricpairs ] gives an explicit description of a moduli space @xmath24 parametrizing @xmath22 and its degeneration .
[ k3coverz23 ] let @xmath16 as in proposition [ mainconstruction ] .
an alternative construction of the @xmath1-cover of @xmath18 branched along @xmath19 is the following .
let @xmath266 .
the restriction to @xmath267 of the map @xmath268 given in an affine patch by @xmath269 is the @xmath270-cover of @xmath18 branched along @xmath19 . to show that @xmath267 is smooth , we first observe that @xmath267 is smooth in the complement of the ramification locus because @xmath18 is smooth .
if @xmath267 is singular at a point in the ramification locus , then one can explicitly compute using the equation of @xmath267 that one of the irreducible components of @xmath14 restricts to @xmath18 giving a reducible curve , which is against our assumptions .
therefore @xmath267 is smooth , implying that @xmath267 is a k3 surface ( @xmath271 by adjunction formula and @xmath272 can be computed using the long exact sequence in cohomology associated to @xmath273 ) .
observe that @xmath267 is invariant under the involution @xmath274 given in an affine patch by @xmath275 .
furthermore , @xmath276 has no fixed points ( these being the torus fixed points of @xmath15 , which are not contained in @xmath267 ) .
this implies that @xmath277 is an enriques surface ( see ( * ? ? ?
* proposition viii.17 ) ) which is the @xmath278-cover of @xmath18 branched along @xmath19 .
let @xmath16 be as in proposition [ mainconstruction ] and denote by @xmath21 the unit cube . then , according to definition [ definitionofstabletoricpair ] , @xmath22 is a stable toric pair of type @xmath23 .
let @xmath24 be the projective coarse moduli space parametrizing these pairs and their degenerations ( see theorem [ modulistabletoricpairs ] ) .
in the next proposition we prove some first properties of @xmath24 , and we can see how the geometry of @xmath24 interacts with the combinatorics of the polyhedral subdivisions of @xmath31 .
[ irreducibilitycoarsemoduliofstabletoricpairs ] let @xmath279 be the unit cube .
then the coarse moduli space @xmath24 is irreducible and has dimension @xmath0 .
the irreducibility of @xmath24 follows from the fact that , if @xmath55 is any polyhedral subdivisions of @xmath31 , then @xmath55 is regular and @xmath280 ( see remark [ combinatorialdescriptionpolarizedstabletoricvariety ] ) . to prove the former ,
we know from ( * ? ? ?
* corollary 2.9 ) that a marked polytope with a nonregular subdivision has a nonregular triangulation .
but all the triangulations of @xmath31 are regular ( see ( * ? ? ?
* theorem 3.2 ) or ( * ? ? ?
* theorem 6.3.10 ) ) .
the proof that @xmath280 can be found in section [ completionproofirreducibilitymqbar ] . for the dimension count , we know that the normalization of @xmath281 is the projective toric variety associated to the secondary polytope @xmath282 ( see @xcite for a complete description of this polytope ) , which has dimension @xmath283 ( see ( * ? ? ?
* chapter 7 , theorem 1.7 ) ) . in conclusion ,
@xmath284 . with reference to section [ formulationviehwegmodulistack ] ,
consider the moduli stack @xmath285 with @xmath286 , @xmath287 ( because we want three pairs of divisors , and we do not distinguish divisors in the same pair ) where @xmath10 is a fixed rational number and @xmath288 ( given @xmath20 as in proposition [ mainconstruction ] , it is easy to compute that @xmath289 ) . for a suitably chosen positive integer @xmath290 depending on @xmath291 and @xmath159 ( which does not need to be specified , see @xcite ) , the viehweg s moduli functor @xmath292 above is coarsely represented by a projective scheme , which we denote by @xmath293 .
this is true because we are working with surface pairs ( @xmath286 ) and our coefficients are strictly greater than @xmath294 ( see ( * ? ? ?
* theorem 1.6.1 , case 2 ) ) .
[ s3actiononksbaspace ] the group @xmath295 has a natural action on the viehweg moduli stack @xmath292 given by permuting the labels of the three divisors @xmath296 .
in particular , we have an induced @xmath295-action on the coarse moduli space @xmath293 .
now we want to identify inside @xmath297 the zariski closed subscheme containing a dense open subset whose points parametrize the stable pairs given by @xmath168 together with our three undistinguished pairs of lines of weight @xmath298 .
the next lemma allows us to do so .
[ importantfamily ] the coarse moduli space @xmath24 contains a dense open subset @xmath299 which is a fine moduli space parametrizing all the stable toric pairs @xmath29 with @xmath300 .
@xmath299 is isomorphic to the dense open subtorus of the toric variety @xmath25 .
the universal family on @xmath299 is of the form @xmath301 .
let @xmath55 be the complex of polytopes given by @xmath21 and its faces .
let @xmath159 be the set of vertices of @xmath21 .
following ( * ? ? ?
* definition 2.6.6 ) , let @xmath302(\bbbk)$ ] be the groupoid of stable toric pairs @xmath22 over @xmath41 with the linearized line bundle @xmath303 in which the arrows are the isomorphisms identical on the torus @xmath44 . by (
* lemma 2.6.7 ) , we have that @xmath302(\bbbk)$ ] is equivalent to the quotient stack @xmath304 $ ] , where @xmath305 represents the space of coefficients of the divisor @xmath252 on @xmath15 and @xmath306 acts on @xmath305 as follows : @xmath307 observe that the quotient @xmath308 exists as a scheme and it is isomorphic to @xmath309 .
the stabilizers of the action @xmath310 are trivial .
it follows that the quotient stack @xmath304 $ ] is finely represented by @xmath308 .
this gives us a dense open subset @xmath299 of @xmath24 isomorphic to @xmath309 with a universal family which can be realized as the quotient of @xmath311 under the @xmath309-action on the coefficients @xmath253 .
[ ksbacompactification ] let @xmath301 be the universal family over @xmath299 constructed in lemma [ importantfamily ] .
denote by @xmath312 the open subset of @xmath299 parametrizing stable toric pairs @xmath22 such that @xmath240 is as in proposition [ mainconstruction ] .
let @xmath313 be the pullback of @xmath314 along the inclusion @xmath315 .
then @xmath316 is a family for the moduli functor @xmath292 , and hence we have an induced morphism @xmath317 .
we define @xmath11 to be the closure of the image of @xmath104 .
[ firstpropertiesofksbaspace ] let @xmath318 be a point in @xmath312 and let @xmath22 be the corresponding stable toric pair .
then the fiber of @xmath319 over @xmath320 is in bijection with the @xmath254-orbit of @xmath22 ( this follows from proposition [ mainconstruction ] ) . in particular
, @xmath104 is a quasi - finite map .
this , together with the fact that @xmath104 is dominant , implies that @xmath321 .
@xmath11 is irreducible because @xmath312 is irreducible . in conclusion
, @xmath11 is a @xmath0-dimensional irreducible projective coarse moduli space whose points in a dense open subset parametrize isomorphism classes of @xmath168 together with the three undistinguished weighted pairs of lines ( and hence enriques surfaces with our choice of divisor after an appropriate @xmath1-cover ) .
let @xmath21 be the unit cube and let @xmath208 be a dvr . let @xmath322 be a family of stable toric pairs of type @xmath23 over @xmath218 . define @xmath323 , @xmath324 and assume that the fiber of @xmath80 over the generic point is isomorphic to @xmath325 where @xmath207 is the field of fractions of @xmath210 .
observe that the pair @xmath20 may be not stable : to start with , the restriction @xmath19 makes no sense if @xmath14 is not @xmath118-cartier ( see proposition [ toricboundarycartier ] ) . in this section
we define a new family @xmath326 which is isomorphic to the original one in the complement of the central fiber , but the new central fiber @xmath32 is such that @xmath33 is a stable pair ( @xmath327 denotes the toric boundary of @xmath328 ) .
the stability of the pair @xmath33 is proved later in section [ wallcrossingstudy ] .
let @xmath208 be a dvr with field of fractions @xmath207 , uniformizing parameter @xmath329 and residue field our fixed base field @xmath41 .
let @xmath44 be a torus over @xmath41 with character lattice @xmath45 .
let @xmath65 be a lattice polytope .
define @xmath330 where , for all @xmath331 , @xmath332 , @xmath333 and @xmath334 .
observe that the association @xmath335 gives a height function @xmath336 .
let @xmath337 be the convex hull of the half - lines @xmath338 , @xmath331 , and let @xmath339 be the cone over @xmath340 with vertex at the origin .
then @xmath341 defines the following @xmath91-graded @xmath210-algebra : @xmath342.\ ] ] observe that @xmath343 is an element of @xmath94-degree @xmath54 .
let @xmath344 be the family of stable toric pairs associated to @xmath98 ( see remark [ algebraicdatafamilyofstabletoricpairs ] ) . if @xmath345 is the generic point of @xmath218 , then @xmath346 , where @xmath133 is the toric variety associated to the polytope @xmath21 .
the central fiber @xmath347 is a stable toric pair whose corresponding polyhedral subdivision of @xmath66 is induced by the height function @xmath341 ( hence , it is a regular subdivision ) .
the equation of @xmath348 is given by @xmath349 for more details about this construction we refer to ( * ? ? ?
* section 2.8 ) .
[ definitionofcornercut ] let @xmath21 be the unit cube .
we call a _ corner cut _ the convex hull in @xmath21 of the points @xmath350 and @xmath351 up to symmetries of @xmath21 .
an example of corner cut can be found in figure [ modificationsubdivision ] on the left .
we call _ apex _ the vertex of a corner cut which is at the intersection of three edges of the cube .
let @xmath55 be a polyhedral subdivision of a lattice polytope @xmath21 .
we denote by @xmath352 the set of @xmath109-dimensional faces in @xmath55 .
[ polyhedralsubdivisionpdot ] let @xmath55 be a polyhedral subdivision of @xmath31 .
we define a polyhedral subdivision @xmath353 of @xmath31 via the following algorithm : * @xmath354 ; * if @xmath355 contains no corner cut , define @xmath356 and stop .
otherwise , go to step ( 3 ) ; * let @xmath357 be a corner cut and let @xmath358 be that unique polytope sharing exactly a facet with @xmath359 .
define @xmath360 then redefine @xmath355 to be the polyhedral subdivision of @xmath21 generated by @xmath361 .
go to step ( 2 ) . in figure [ modificationsubdivision ]
we give an explicit example of @xmath353 given @xmath55 . of the polyhedral subdivision @xmath55 ] let us adopt the same notation introduced in section [ familiesofstabletoricpairsoveradvr ] , but set @xmath21 to be the unit cube .
we define another @xmath91-graded @xmath210-algebra induced by @xmath95 as follows .
denote by @xmath55 the regular polyhedral subdivision of @xmath31 associated to @xmath347 .
assume @xmath363 is a corner cut and let @xmath364 be that unique polytope sharing exactly a facet with @xmath359 .
denote by @xmath365 the unique hyperplane in @xmath366 containing the points @xmath367 for @xmath368 .
if @xmath369 is the apex of the corner cut @xmath359 , then there exists a unique positive rational number @xmath370 such that @xmath371 .
moreover , up to a ramified finite base change , we can assume that @xmath370 is integral .
let us consider the height function @xmath372 define a new
@xmath91-graded @xmath210-algebra @xmath373 as we did in section [ familiesofstabletoricpairsoveradvr ] , but using @xmath374 in place of @xmath341 .
observe that @xmath375 is a degree preserving embedding of graded algebras .
therefore @xmath376 is an element of @xmath94-degree @xmath54 and the pair @xmath377 corresponds to a family @xmath378 of stable toric varieties together with a cartier divisor @xmath379 given by the vanishing of @xmath95 . observe that @xmath380 and @xmath381 are isomorphic over @xmath213 by construction .
the central fiber @xmath382 is not a stable toric pair if and only if @xmath55 contains corner cuts .
however , @xmath383 is a polarized stable toric variety whose corresponding polihedral subdivision of @xmath31 is @xmath353 ( see definition [ polyhedralsubdivisionpdot ] ) .
the equation of @xmath384 is given by @xmath385 observe that if @xmath55 has no corner cuts , then @xmath386 . [ independentfromthechoiceofthefamily ] with the same notation introduced above , denote by @xmath29 ( resp .
@xmath32 ) the central fiber of @xmath322 ( resp .
@xmath326 ) .
then @xmath32 only depends on @xmath29 and it is independent from the whole family @xmath344 .
[ modificationcentralfiberbullet ] let @xmath29 be a stable toric pair of type @xmath23 .
then define @xmath32 to be the central fiber of @xmath326 where @xmath322 is any one - parameter family with central fiber @xmath29 and smooth generic fiber ( there exists such family @xmath322 because @xmath24 is irreducible , hence @xmath29 is smoothable ) .
@xmath32 is well defined by remark [ independentfromthechoiceofthefamily ] .
observe that , if @xmath55 has no corner cuts , then @xmath387 .
throughout this section let @xmath21 be the unit cube . we show that if @xmath29 is a stable toric pair of type @xmath23 , then @xmath388 is stable . before formally stating this result
, we need some preliminaries . consider a stable toric pair @xmath29 of type @xmath23 and let @xmath55 be the corresponding polyhedral subdivision of @xmath31 .
let @xmath389 be the normalization of @xmath30 , where @xmath390 is the toric variety corresponding to the polytope @xmath359 .
then we denote by @xmath391 the toric boundary of @xmath390 , by @xmath392 the conductor divisor and by @xmath393 the restriction to @xmath390 of the preimage of @xmath18 under the normalization morphism .
define @xmath394 and @xmath395 analogously with @xmath32 in place of @xmath29 .
[ toricboundarycartier ] let @xmath29 be a stable toric pair of type @xmath23 and let @xmath55 be the associated polyhedral subdivision of @xmath31 .
if @xmath55 does not contain a corner cut , then @xmath14 is cartier .
if @xmath55 contains a corner cut , then @xmath14 is not @xmath118-cartier .
define a piecewise linear function on the normal fan of each maximal dimensional polytope @xmath57 as follows : associate to the @xmath94-generators of the rays , negative the lattice distance between the facet of @xmath396 normal to the ray and the lattice point @xmath17 .
observe that this number is @xmath13 if the facet contains @xmath17 , and @xmath397 otherwise .
if @xmath55 has no corner cuts , then this gives a cartier divisor on @xmath30 equal to @xmath14 .
now assume @xmath55 contains a corner cut @xmath359 .
denote with @xmath96 that unique polytope in @xmath55 sharing exactly a facet with @xmath359 .
let @xmath398 be the three edges of @xmath359 which do not contain the apex .
observe that @xmath399 can be contained in two or three maximal dimensional polytopes in @xmath55 ( @xmath359 and @xmath96 included ) , @xmath152 .
if some @xmath399 is contained in three maximal dimensional polytopes , then take a point @xmath400 lying on the torus invariant line corresponding to @xmath399 .
if @xmath14 is @xmath118-cartier , then @xmath401 is given by the vanishing of one equation in an open neighborhood of @xmath402 for some @xmath403 .
however , the vanishing locus of this equation on @xmath404 has codimension @xmath132 , which can not be .
assume that each @xmath399 is only contained in @xmath359 and @xmath96 .
denote by @xmath113 the normalization . if @xmath14 is @xmath118-cartier , then @xmath405 is also @xmath118-cartier .
but this is a contradiction because there is no @xmath118-piecewise linear function on the normal fan of @xmath96 corresponding to @xmath406 ( consider the normal cone to a vertex of @xmath96 in common with @xmath359 ) .
[ stabilityofthemodifiedpair ] let @xmath21 be the unit cube and let @xmath29 be a stable toric pair of type @xmath23 . consider @xmath32 as in definition [ modificationcentralfiberbullet ] .
then @xmath33 is a stable pair ( it makes sense to consider @xmath407 by proposition [ toricboundarycartier ] ) .
let @xmath55 be the polyhedral subdivision of @xmath31 associated to @xmath29 .
we show that for all @xmath408 , the pair @xmath409 is stable .
it is easy to observe that there are four possibilities for @xmath359 up to symmetries of @xmath21 as shown in figure [ polytopeswithoutcornercuts ] .
[ definitiontypepolytopeinpbullet ] we say that @xmath408 has type @xmath410 ( resp .
@xmath411 ) if @xmath359 is equal to the polytope in figure [ polytopeswithoutcornercuts](a ) ( resp .
@xmath411 ) up to symmetries of @xmath21 . up to symmetries of @xmath21 ] [ adescription ] given @xmath359 of type @xmath410 , then the pair @xmath409 is stable .
@xmath413,\\ & d_p^\bullet = v(w_0)+v(w_1),\\ & \delta_p^\bullet - d_p^\bullet = v(w_2)+v(w_3),\\ & b_p^\bullet = v(a_0w_0+\ldots+a_3w_3).\end{aligned}\ ] ] where @xmath414 for all @xmath415 ( because there is no corner cut that can possibly be contained in @xmath359 ) .
then @xmath395 is isomorphic to @xmath2 and @xmath416 restricts to @xmath395 giving four lines in general linear position , implying that @xmath409 is log canonical .
finally , if @xmath365 denotes a generic line in @xmath395 , then @xmath417 which is ample .
the next lemma is used in the analysis of the cases @xmath359 of type @xmath418 and @xmath419 .
[ threeincidentlinesweightslogcanonical ] let @xmath420 be three distinct lines in @xmath421 through the point @xmath422 .
then the pair @xmath423 is log canonical .
consider the blow up morphism @xmath424 and let @xmath425 be the exceptional divisor .
then we have that @xmath426 this implies that @xmath423 is log canonical because @xmath427 is a log resolution of singularities .
[ bdescription ] given @xmath359 of type @xmath428 , then the pair @xmath409 is stable .
@xmath429,\\ & d_p^\bullet = v(w_0,w_2)+v(w_0,w_3)=v(w_0w_1-w_2w_3,w_0),\\ & \delta_p^\bullet - d_p^\bullet = v(w_1,w_2)+v(w_1,w_3)+v(w_0w_1-w_2w_3,w_4)\\ & \hspace{1.6cm}=v(w_0w_1-w_2w_3,w_1)+v(w_0w_1-w_2w_3,w_4),\\ & b_p^\bullet = v(a_0w_0+\ldots+a_4w_4)\cap x_p^\bullet,\end{aligned}\ ] ] where @xmath414 for @xmath430 and @xmath431 can possibly vanish ( we can have at most one corner cut contained in @xmath359 ) .
@xmath432 because it is a hyperplane section of the projective cone @xmath433 which does not pass through the vertex .
now let us study the restriction of @xmath434 and @xmath435 to @xmath395 .
this boils down to understand how the coordinate hyperplanes @xmath436 , @xmath437 restrict to @xmath395 .
first of all , observe that @xmath438 cuts on @xmath395 a curve @xmath159 of divisor class @xmath439 . to show this , denote by @xmath440 the divisor class of @xmath441 .
if @xmath442 and @xmath143 is a generic hyperplane in @xmath443 , then the self - intersection of @xmath159 is given by @xmath444 on the other hand , @xmath445 , implying that @xmath446 .
for @xmath447 , @xmath448 is always reducible , so it is given by two curves of divisor classes @xmath449 and @xmath450 . if @xmath451 , then it is easy to check that @xmath452 is smooth for a generic choice of the coefficients , but it can possibly break into two curves . in any case , @xmath416 restricts to @xmath395 giving a simple normal crossing divisor , implying that @xmath409 is log canonical .
if @xmath453 , then @xmath452 is automatically irreducible , but it passes through the singular point of @xmath454 . however , we can still conclude that @xmath409 is log canonical by lemma [ threeincidentlinesweightslogcanonical ] .
finally , observe that @xmath455 which is ample .
in the cases @xmath359 of type @xmath456 or type @xmath419 , the ideas are similar to the ones we already used in the previous propositions . therefore we discuss these remaining cases in the appendix ( see propositions [ cdescription ] and [ ddescription ] ) . [ explicitdescriptionirreduciblecomponentsdegenerations ] from the proof of theorem [ stabilityofthemodifiedpair ] we have an explicit description of all the possible @xmath457 for any stable toric pair @xmath29 of type @xmath23 . here
we want to summarize these possibilities . in figure
[ explicitdegenerationsinvariantwallcrossing ] , a triangle ( resp .
trapezoid , parallelogram ) means @xmath458 ( resp .
@xmath459 , @xmath460 ) .
@xmath461 is represented by the thickened segments and @xmath462 by the colored segments .
first , let us assume that @xmath55 has no corner cuts , so that @xmath387 . in all the cases that follow , @xmath463 is simple normal crossing .
* @xmath464 and @xmath465 ( resp .
@xmath466 ) consists of two lines ; * @xmath467 and @xmath465 consists of two incident rulings .
@xmath466 is given by two incident rulings and a curve of divisor class @xmath439 which can possibly be reducible ; * @xmath468 and @xmath465 is a line disjoint from the exceptional divisor .
@xmath466 is given by two fibers and two lines disjoint from the exceptional divisor ; * as above , but @xmath465 is the union of the exceptional divisor and a fiber ; * @xmath393 is isomorphic to the union of @xmath2 and @xmath460 glued along a line in @xmath2 and a ruling in @xmath460 .
@xmath465 consists of a line in @xmath2 and a ruling in @xmath460 .
@xmath466 is given by two lines on the @xmath2 component and four rulings on @xmath460 arranged as shown in figure [ explicitdegenerationsinvariantwallcrossing ] ; * @xmath469 and @xmath470 .
@xmath463 is as in proposition [ d16asz22coverandconverse ] ( this is the generic case ) , or some lines can possibly break into two intersecting @xmath178-curves ; * @xmath393 is a singular del pezzo surface of degree @xmath471 with exactly one @xmath472 singularity .
this singularity can lie on at most one irreducible component of @xmath463 ; * @xmath393 is isomorphic to the union of two copies of @xmath460 glued along a ruling and an irreducible curve of divisor class @xmath439 .
@xmath463 is given by four rulings on one component and six rulings on the other .
these are arranged as shown in figure [ explicitdegenerationsinvariantwallcrossing ] ; * @xmath393 is isomorphic to the union of three copies of @xmath460 glued along rulings as shown in figure [ explicitdegenerationsinvariantwallcrossing ] .
@xmath463 consists of four rulings on each component as shown in figure [ explicitdegenerationsinvariantwallcrossing ] .
now assume that @xmath55 has a corner cut and let @xmath408 . in this case , the possibilities for the pair @xmath409 are as above , with the difference that now @xmath462 is allowed to have triple intersection points . for @xmath408 for a generic choice of the coefficients as explained in remark [ explicitdescriptionirreduciblecomponentsdegenerations ] .
the restriction @xmath462 is in blue , green and red , @xmath461 is thickened and , if two irreducible components are glued together , then the gluing curve is in yellow ]
in theorem [ allyouneedtoknowaboutksbapoints ] we describe a surjective ( algebraic ) morphism @xmath39 which gives a geometric interpretation of all the @xmath41-points of @xmath11 ( the global geography of @xmath11 is described in section [ globalgeography ] ) .
we need some preliminaries from @xcite .
consider a map @xmath473 , where @xmath207 is the field of fractions of a dvr @xmath208 and @xmath133 is a proper scheme . by the valuative criterion of properness
, @xmath474 extends in a unique way to a map @xmath475 .
we denote by @xmath476 the point @xmath477 .
[ lemmaextensionmapmodulispaces ] suppose @xmath478 and @xmath479 are proper schemes over a noetherian scheme @xmath7 with @xmath478 normal .
let @xmath480 be an open dense set and @xmath481 an @xmath7-morphism .
then @xmath104 extends to an @xmath7-morphism @xmath482 if , and only if , for any dvr @xmath208 and any morphism @xmath483 , the point @xmath484 of @xmath479 is uniquely determined by the point @xmath476 of @xmath478 . in what follows ,
given a scheme @xmath30 , we denote by @xmath485 the normalization of @xmath30 with the reduced scheme structure .
[ allyouneedtoknowaboutksbapoints ] let @xmath21 be the unit cube and denote by @xmath25 the normalization of @xmath24 with the reduced scheme structure
. then there is a surjective ( algebraic ) morphism @xmath39 which is given on @xmath41-points by @xmath486 set @xmath487 ( which was constructed in definition [ ksbacompactification ] ) and let @xmath488 .
the claimed morphism @xmath489 is obtained by extending @xmath490 to the whole @xmath478 by means of theorem [ lemmaextensionmapmodulispaces ] .
let @xmath208 be any dvr and consider a map @xmath491 .
the point @xmath492 corresponds to a stable toric pair @xmath29 of type @xmath23 .
denote by @xmath14 the toric boundary of @xmath30 .
if we prove that @xmath484 corresponds to the stable pair @xmath33 , then we are done by theorem [ lemmaextensionmapmodulispaces ] , because this shows that @xmath484 only depends on @xmath476 , and not from the way we approach it ( see remark [ independentfromthechoiceofthefamily ] ) .
let @xmath493 be as in definition [ ksbacompactification ] .
let @xmath494 be the pullback of @xmath495 under the map @xmath474 and denote by @xmath322 its completion over @xmath218 ( or a finite ramified base change of it ) .
consider @xmath326 and , if @xmath116 denotes the toric boundary of @xmath15 , let @xmath496 be the closure of @xmath497 in @xmath498 .
then the central fiber of @xmath499 , which is @xmath33 by construction , is the stable pair corresponding to @xmath484 .
[ propertiesofthemapstptoksba ] let @xmath39 be the morphism in theorem [ allyouneedtoknowaboutksbapoints ] .
then the induced morphism @xmath500 is an isomorphism .
observe that @xmath254 acts on @xmath25 because @xmath25 is the projective toric variety associated to the secondary polytope of @xmath31 .
the modular interpretation of this action is the following : @xmath254 acts on the stable toric pair @xmath29 by changing the torus action on it .
in particular , it is clear that the morphism @xmath39 is @xmath254-equivariant .
therefore , we have an induced morphism @xmath501 which lifts to the normalization of @xmath11 .
we already observed that the fibers of the restriction of @xmath39 to @xmath312 are exactly @xmath254-orbits ( see definition [ ksbacompactification ] and observation [ firstpropertiesofksbaspace ] ) .
this implies that the morphism @xmath502 is generically bijective .
then the fact that @xmath502 is an isomorphism follows from the zariski main theorem ( see ( * ? ? ? * chapter iii , section 9 ) ) after we show that @xmath39 is quasi - finite . it will be enough to prove that no @xmath54-dimensional boundary stratum of @xmath25 is contracted by @xmath39 .
these strata are in correspondence with the minimal elements of the poset of regular polyhedral subdivisions of @xmath31 which are not triangulations ( all the polyhedral subdivisions of @xmath503 are regular , see the proof of proposition [ irreducibilitycoarsemoduliofstabletoricpairs ] ) .
if @xmath55 is one of these polyhedral subdivisions , then it has to contain a subpolytope of @xmath21 with vertices in @xmath504 which can be subdivided further only once .
an easy enumeration shows that @xmath55 has to contain one of the polytopes listed in figure [ threepolytopesinonedimensionalsubdivision ] .
denote by @xmath359 one of such polytopes and let @xmath29 be any stable toric pair with @xmath55 as corresponding polyhedral subdivision of @xmath31 .
then , as @xmath29 varies among the stable toric pairs parametrized by the @xmath54-dimensional boundary stratum corresponding to @xmath55 , @xmath505 describes a @xmath54-dimensional family of stable pairs . to prove this ,
if @xmath506 is the polytope corresponding to @xmath57 , it is enough to look at the irreducible component of @xmath505 corresponding to @xmath507 .
which can be subdivided further only once ( see the proof of theorem [ propertiesofthemapstptoksba ] ) ] we define the _ boundary _ of the moduli space @xmath36 to be the closed subset whose @xmath41-points parametrize stable pairs @xmath508 where @xmath18 is reducible .
a _ stratum _ of the moduli space @xmath11 is an irreducible component of the boundary or the intersection of some of these irreducible components . [ combinatoricsstratificationstatement ] the boundary of @xmath36 is stratified as shown in figure [ explicitstratification ] , where the strata are organized from bottom to top in increasing order of dimension . going from left to right in the same diagram , call the @xmath13-dimensional strata even , odd of type @xmath54 and odd of type @xmath132 ( this terminology is borrowed from @xcite ) .
then the strata specializing to the even @xmath13-dimensional stratum correspond bijectively to the polyhedral subdivisions of @xmath31 without corner cuts up to symmetries of @xmath21 . by theorem [ allyouneedtoknowaboutksbapoints ]
any stable pair parametrized by @xmath11 is in the form @xmath33 for some stable toric pair @xmath29 parametrized by @xmath25 . in remark [ explicitdescriptionirreduciblecomponentsdegenerations ]
we listed all the possibilities for @xmath409 , and these pairs can be glued together for @xmath408 to recover @xmath33 , where @xmath353 is a polyhedral subdivisions of @xmath31 without corner cuts .
these subdivisions are shown in figure [ subdivisionscubewithoutcornercuts ] up to symmetries of @xmath21 .
the end result is shown in figure [ explicitstratification ] , where for each stratum we chose the generic element to represent it .
the three colors ( blue , green and red ) used to draw the divisor have the following meaning : lines sharing the same color come from the same pair of lines on the original @xmath168 .
we want to remark that even though we use three different colors , we do not distinguish the three pairs because we took the quotient by @xmath295 in our definition of @xmath11 .
the yellow color is used to indicate lines along which two irreducible components are glued .
the claimed combinatorial interpretation of the strata specializing to the even @xmath13-dimensional stratum follows immediately after comparing figure [ subdivisionscubewithoutcornercuts ] and figure [ explicitstratification ] . without corner
cuts up to symmetry and ordered by refinement ] ] after describing all the degenerations of @xmath168 together with the three undistinguished pairs of weighted lines parametrized by @xmath11 , we want to discuss what are the overlying @xmath1-covers , which correspond to actual degenerations of enriques surfaces .
we describe such @xmath1-cover for the stable pairs @xmath508 parametrized by a general point on each boundary stratum of @xmath11 .
* consider the stable pair parametrized by the odd @xmath13-dimensional stratum of type @xmath132 .
then its @xmath1-cover is isomorphic to the quotient of @xmath509 by the involution @xmath510 ( see remark [ k3coverz23 ] and proof of proposition [ ddescription ] ) , which is given by three copies of @xmath460 glued along rulings in such a way that its dual complex gives a triangulation of the real projective plane . *
the stable pair parametrized by a general point on the maximal @xmath54-dimensional stratum of @xmath11 has @xmath1-cover isomorphic to the quotient of @xmath511 , @xmath512 , by @xmath510 ( again , see remark [ k3coverz23 ] and proof of proposition [ ddescription ] ) .
if @xmath513 , then its irreducible components are a copy of @xmath460 and an elliptic fibration @xmath514 over @xmath149 with fibers isomorphic to @xmath425 and two double fibers .
these two surfaces are glued along @xmath515 and a fiber of @xmath514 which is not double . *
consider the stable pair parametrized by a general point on the irreducible boundary divisor specializing to two @xmath13-dimensional strata .
let us describe the @xmath1-cover @xmath30 of one of the two irreducible components , which are isomorphic to @xmath516 .
denote by @xmath341 ( resp .
@xmath104 ) a section of self - intersection @xmath54 ( resp . a fiber ) .
then the building data for the cover @xmath517 is given by @xmath518 . from this
we obtain that the canonical class of @xmath30 is given by @xmath519 moreover , @xmath520 is big , nef and @xmath521 , implying that @xmath30 is a weak del pezzo surface of degree @xmath54 .
the total degeneration is given by two of such weak del pezzo surface glued along a genus 1 curve .
* let @xmath30 be the @xmath1-cover of the @xmath460 in figure [ explicitdegenerationsinvariantwallcrossing](b ) .
denote by @xmath522 and @xmath523 two incident rulings .
then the building data for the cover @xmath524 is given by @xmath525 .
from this we obtain that @xmath526 .
it follows that @xmath520 is ample and @xmath527 .
hence , @xmath30 is a del pezzo surface of degree @xmath132 .
* let us describe the @xmath1-cover of the @xmath2 in figure [ explicitdegenerationsinvariantwallcrossing](a ) .
if @xmath528 denotes a line in @xmath2 , then the building data for the cover @xmath529 is given by @xmath530 .
therefore , @xmath531 , implying that @xmath520 is ample and @xmath532 . hence , @xmath30 is a del pezzo surface of degree @xmath0 .
* let us describe the @xmath1-cover of the @xmath460-component in figure [ explicitdegenerationsinvariantwallcrossing](c3 ) . if @xmath522 and @xmath523 denote two incident rulings , then the building data for the cover @xmath524 is given by @xmath533 .
therefore , @xmath534 , @xmath520 is ample , @xmath532 and @xmath30 is a del pezzo surface of degree @xmath0 .
in what follows we construct a morphism from the ksba compactification @xmath36 to the baily - borel compactification of @xmath535 , where @xmath35 is the period domain parametrizing @xmath5-polarized enriques surfaces ( details in section [ relationksbabailyborel ] ) . section [ generalizedtypestudy ] contains a technical result which is fundamental to construct such morphism . in section [ semitoriccompactification ]
we show that the behavior of the boundary strata of @xmath36 specializing to the even @xmath13-dimensional cusp has another combinatorial interpretation in terms of vinberg diagrams .
let @xmath536 be our base field .
in section [ relationwithbailyborel ] our focus moves from enriques surfaces to k3 surfaces .
the reason for this is that in theorem [ morphismtobailyborelcompactification ] we want to compute limits of one - parameter families of @xmath5-polarized enriques surfaces in the baily - borel compactification of @xmath535 , and this is done by looking at their corresponding k3 covers .
let @xmath14 be the unit disk @xmath537 and let @xmath538 .
we are interested in proper flat families @xmath539 with @xmath540 smooth and where @xmath541 is a smooth k3 surface for all @xmath542 .
equip the family @xmath540 with an effective relative cartier divisor @xmath543 such that @xmath544 is a family of stable pairs .
let @xmath80 be a semistable degeneration with @xmath545 ( or simply kulikov degeneration for short ) completing @xmath540 over @xmath14 ( see @xcite and @xcite ) .
recall that the central fiber @xmath546 can be of type i , ii or iii ( see ( * ? ? ?
* theorem ii ) ) . in type ii , denote by @xmath547 the @xmath548-invariant of one of the mutually isomorphic genus @xmath54 double curves on @xmath546 . then define @xmath549 to be the closure of @xmath543 inside @xmath80 ( @xmath549 is flat over @xmath14 ) .
we can define a second completion of @xmath550 over @xmath14 , which we denote by @xmath551 , such that @xmath552 is a stable pair for all @xmath10 . with the notation introduced above ,
define the _ dual graph of @xmath553 _ as follows . draw a vertex @xmath554 for each irreducible component @xmath555 of @xmath553 .
then , given any two distinct irreducible components @xmath555 and @xmath556 , draw one edge between @xmath554 and @xmath557 for each irreducible curve in @xmath558 . if an irreducible component @xmath555 self - intersects along a curve @xmath159 , then draw one loop on @xmath554 for each irreducible component of @xmath159 . denote by @xmath559 the dual graph of @xmath553 .
[ ksbatype ] let @xmath560 as defined above .
we say that @xmath553 has _
generalized type i , ii _ or _ iii _ if the following hold : * type i : @xmath559 consists of one vertex and @xmath553 has at worst du val singularities ; * type ii : @xmath559 is a chain .
if there is only one vertex , then @xmath553 has an elliptic singularity . if there are at least two vertices and the double curves are mutually isomorphic genus @xmath54 curves , then denote by @xmath561 the @xmath548-invariant of one of these ; * type iii : @xmath559 is obtained by contracting the @xmath54-skeleton of a triangulation of a sphere . if @xmath559 consists of only one vertex , then @xmath553 has worse than an elliptic singularity .
the proof of the following theorem , which builds upon the proof of ( * ? ? ?
* theorem 2.9 ) , was communicated to me by valery alexeev .
[ goodbehaviortypeofdegeneration ] with the notation introduced above , @xmath546 and @xmath553 have the same type .
in addition , if @xmath562 have type ii and @xmath561 can be defined , then @xmath563 . the proof of ( * ? ? ? *
theorem 2.9 ) describes a procedure to construct the unique stable model @xmath564 by modifying @xmath565 .
this procedure consists mainly of the following two steps : * step @xmath54 : replace @xmath80 with another kulikov degeneration such that @xmath549 is nef and it does not contain double curves or triple points ( this may involve , among other things , base changes ) ; * step @xmath132 : for @xmath566 , the line bundle @xmath567 induces a birational morphism @xmath568 where @xmath569 ( see ( * ? ? ? * theorem 2 , part ( i ) ) ) .
this birational morphism consists of the contraction of some components or curves in the central fiber @xmath546 . in step @xmath54
, the new kulikov model is obtained from the one we started by applying elementary modifications of type @xmath13 , i and ii ( see @xcite for their definitions ) , base changes , and blow ups of @xmath546 along double curves or triple points .
the elementary modifications and the blow ups obviously do not change the type of the central fiber . a description of how the central fiber is modified after a base change can be found in @xcite , and also in this case the type does not change . in step @xmath132 , it follows from our definition of generalized type that @xmath553 has the same type as @xmath546 .
this shows that if @xmath546 has type i , ii or iii , then @xmath553 has type i , ii or iii respectively .
the converse follows from this and from the uniqueness of the stable model .
the claim about the @xmath548-invariants also follows from our discussion .
the period domain @xmath570 parametrizing @xmath5-polarized enriques surfaces ( where @xmath365 is the lattice @xmath571 , see ( * ? ? ?
* section 3.4.2 ) for the details ) and its baily - borel compactification @xmath34 ( @xmath572 denotes the isometry group of @xmath365 ) are studied in @xcite .
oudompheng shows that the boundary of @xmath34 consists of two rational @xmath54-cusps and three @xmath13-cusps ( see ( * ? ? ?
* proposition 7.7 ) ) . in what follows
we construct a birational morphism @xmath37 which highlights a correspondence between the boundary strata of the two compactifications .
the two rational @xmath54-cusps @xmath573 ( the notation is borrowed from @xcite ) of the baily - borel compactification @xmath34 are degree @xmath12 covers of the modular curve @xmath574 ( see ( * ? ? ?
* section 7.3 ) ) .
[ foldingcompactification ] there exists a compactification @xmath575 of @xmath535 obtained from @xmath34 by gluing the three @xmath13-cusps together and by gluing the two @xmath54-cusps to a rational curve with one node whose smooth points correspond to isomorphism classes of elliptic curves .
moreover , @xmath34 is the normalization of @xmath575 .
the main technique we use for the proof is ( * ? ? ?
* theorem 5.4 ) .
let us briefly recall it .
let @xmath576 be a scheme , @xmath577 a closed subscheme of @xmath576 and @xmath578 a finite morphism .
consider the ringed space @xmath579 and the cocartesian square ( a)[matrix of math nodes , row sep=2em , column sep=2em , text height=1.5ex , text depth=0.25ex ] y&y + x&x .
+ ; ( a-1 - 1 ) edge node[above](a-1 - 2 ) ; ( a-1 - 1 ) edge node[left](a-2 - 1 ) ; ( a-2 - 1 ) edge node[below](a-2 - 2 ) ; ( a-1 - 2 ) edge node[right](a-2 - 2 ) ; let us assume that any finite sets of points in @xmath576 ( resp .
@xmath133 ) are contained in an open affine subset of @xmath576 ( resp .
@xmath133 ) . then @xmath30 is a scheme verifying the same property on finite sets of points , the above diagram is cartesian
, @xmath580 is a closed immersion , the morphism @xmath581 is finite and it induces an isomorphism @xmath582 . in our case , we want to first consider the finite morphism @xmath583 to build the projective scheme @xmath584 ( the hypothesis on finite sets of points is satisfied because @xmath34 and @xmath585 are projective ) .
repeat the gluing on @xmath478 by considering now @xmath586 to obtain @xmath479 .
now glue together the two copies of @xmath585 on @xmath479 to obtain @xmath587 .
finally , identify the images in @xmath587 of the three @xmath13-cusps to obtain the claimed compactification @xmath575 .
the isomorphism @xmath588 follows from the zariski main theorem because @xmath34 is normal and the morphism @xmath589 is finite and birational .
[ familiesofenriquesandk3s ] recall the family @xmath590 in definition [ ksbacompactification ] .
we denote by @xmath591 ( resp .
@xmath592 ) the @xmath1-cover ( res .
@xmath270-cover ) of @xmath313 branched along @xmath593 , which is a family of @xmath5-polarized enriques surfaces ( resp .
k3 surfaces ) .
observe that @xmath592 , like @xmath313 , is embedded in @xmath594 .
[ morphismtobailyborelcompactification ] there exists a birational morphism @xmath37 which maps the boundary of @xmath36 to the boundary of @xmath34 .
the open subset @xmath595 carries the family @xmath591 of @xmath5-polarized enriques surfaces and it induces a morphism @xmath596 whose fibers are @xmath254-orbits .
we show that this morphism extends to the whole @xmath25 giving a @xmath254-equivariant morphism .
this extension is induced by the universal property of the normalization after we extend the composition @xmath597 , which we denote by @xmath598 , to the whole @xmath25 ( the compactification @xmath575 is constructed in lemma [ foldingcompactification ] ) .
to extend @xmath598 we use again theorem [ lemmaextensionmapmodulispaces ] .
so let @xmath207 be the field of fractions of a dvr @xmath208 and consider any @xmath491 .
we show that @xmath599 can be computed using only @xmath600 .
let @xmath29 be the stable toric pair parametrized by @xmath600 and consider the corresponding stable pair @xmath33 .
we distinguish the following three cases : * case i : @xmath601 is irreducible ; * case ii : @xmath601 has exactly two irreducible components ; * case iii : @xmath601 has at least three irreducible components .
denote by @xmath602 ( resp .
@xmath603 ) the image of the @xmath13-cusps ( resp .
@xmath54-cusps ) under the morphism @xmath589 .
the point @xmath604 can be computed as follows .
@xmath312 carries the family of @xmath605 surfaces @xmath606 , where @xmath14 is the toric boundary of @xmath15 .
let @xmath607 be the completion over @xmath218 ( or a finite ramified base change of it ) of the restriction of @xmath592 to @xmath213 ( we omit the divisor for simplicity of notation ) .
in particular , @xmath608 is the @xmath270-cover of @xmath601 branched along @xmath609 and it is the stable model of a degeneration of smooth k3 surface pairs .
let @xmath610 be a birational modification of @xmath607 which is a kulikov degeneration .
then @xmath611 determines a unique point in @xmath575 , which depends on the type of @xmath611 as follows : * type i : @xmath612 is the image under the quotient @xmath613 of the period point corresponding to @xmath611 ; * type ii : @xmath614 and it corresponds to @xmath615 ; * type iii : @xmath616 .
it is easy to observe that @xmath601 falls into case i ( resp .
ii , iii ) if and only if @xmath608 has generalized type i ( resp .
ii , iii ) .
moreover , the generalized type of @xmath608 equals the type of @xmath611 by theorem [ goodbehaviortypeofdegeneration ] . in addition
, if we are in case ii , then it makes sense to define the @xmath548-invariant @xmath617 , and this equals @xmath615 again by theorem [ goodbehaviortypeofdegeneration ] . in conclusion , we proved that @xmath604 only depends on @xmath476 .
[ correspondenceoflabelsfor0strata ] in @xcite the three @xmath13-cusps are called even , odd of type @xmath54 and odd of type @xmath132 . from the way we constructed the morphism @xmath37 ,
it is not clear to which one of these @xmath13-cusps is a given @xmath13-stratum mapped to .
let us briefly show that a @xmath13-dimensional boundary stratum maps to the @xmath13-cusp with the same label ( recall the terminology for the @xmath13-dimensional boundary strata in theorem [ combinatoricsstratificationstatement ] ) .
it is enough to show that a given point in the interior of the maximal @xmath54-dimensional stratum of @xmath36 is mapped to the @xmath54-cusp @xmath618 .
consider a smooth one - parameter family with fibers isomorphic to @xmath168 .
equip this family with a divisor of coefficient @xmath298 cutting on each fiber our usual configuration of three pairs of lines without triple points , exception made for the central fiber where two lines belonging to the same pair come together and the other four lines are general .
it is easy to compute that the limit of this one - parameter family in @xmath36 lies on the maximal @xmath54-dimensional stratum ( to show this , one has to compute the stable replacement of the central fiber , but this is done by blowing up the double line and then contracting the strict transforms of the two @xmath178-curves intersecting the double line ) .
the limit of the same family in @xmath34 belongs to @xmath618 , and this can be deduced using the git interpretation of @xmath34 in terms of degenerations of line arrangements in @xmath168 ( see @xcite ) .
this shows what we claimed .
the even @xmath13-cusp of @xmath34 corresponds to the isometry class of a primitive isotropic vector @xmath619 such that @xmath620 , which denotes the unique unimodular lattice of signature @xmath621 ( see ( * ? ? ?
* section 7 ) ) .
let @xmath622 be the discrete reflection subgroup of @xmath623 generated by the reflections with respect to the vectors of square @xmath397 .
running vinberg s algorithm to determine a fundamental domain for @xmath622 ( see ( * ? ? ?
* section 1 ) and ( * ? ? ? * theorem 2.6 bis ) for the stopping condition ) , we obtain the _ coxeter diagram _ ( see ( * ? ? ?
* section 1 ) or ( * ? ? ?
* table 2.2.8 ) ) shown on the left in figure [ coxeterdiagramandsubdiagrams ] . for the computations , we chose @xmath624 as our initial vector .
now observe that there is a bijection between the strata of the boundary of @xmath36 specializing to the even @xmath13-dimensional stratum ( see figure [ explicitstratification ] ) and the elliptic and maximal parabolic subdiagrams of the coxeter diagram up to symmetries of the hexagon ( on the right in figure [ coxeterdiagramandsubdiagrams ] ) .
therefore , the compactification @xmath36 is toroidal at the even @xmath13-dimensional stratum .
similar computations for the odd @xmath13-cusps of the baily - borel compactification did not give any significant interpretation for the remaining part of the boundary of @xmath36 . however , @xmath36 is isomorphic to the baily - borel compactification in a neighborhood of the odd @xmath13-dimensional stratum of type @xmath132 .
these considerations show that @xmath36 is one of the compactifications studied in @xcite ( we do not elaborate on the details of it though ) .
and on the right its elliptic and maximal parabolic subdiagrams up to symmetries ] in the case considered in this paper , it is remarkable that the theory of stable pairs produced a compactification @xmath36 which is toroidal at the even @xmath13-dimensional stratum , baily - borel at the odd @xmath13-dimensional stratum of type @xmath132 and a mixture " of the two compactifications at the odd @xmath13-dimensional stratum of type @xmath54 .
in particular , this illustrates what is the behavior that should be expected in general when using stable pairs to compactify moduli spaces .
let @xmath55 be a complex of equidimensional polytopes together with a reference map @xmath56 .
let @xmath625 be the set of maximal dimensional polytopes in @xmath55 and recall that @xmath626 ( which we shorten by @xmath627 ) denotes the saturated sublattice of @xmath59 generated by @xmath628 .
for any @xmath629 , denote by @xmath630 the intersection @xmath631 .
now , an element of @xmath64 is a tuple @xmath632 indexed by pairs in @xmath236 , with @xmath633 satisfying the _ cocycle condition _ and _ modulo coboundaries_. the cocycle condition is the following : if @xmath634 is any triple , then we must have @xmath635 .
a coboundary is a tuple in the form @xmath636 for @xmath637 and @xmath638 .
therefore , we have an equivalence relation on the cocycles given by @xmath639 [ removehangingpolytope ] let @xmath55 be a connected complex of polytopes obtained by gluing finitely many polytopes of the same dimension .
let @xmath359 be one of these polytopes and assume @xmath359 has exactly one facet in common with some other maximal dimensional polytope in @xmath55 . let @xmath640 be the complex of polytopes obtained from @xmath55 by eliminating @xmath359
. then @xmath641 .
label @xmath359 by @xmath54 and the adjacent one by @xmath132 .
consider an arbitrary cocycle @xmath642 .
let @xmath643 be such that @xmath644 and consider the coboundary @xmath645 ( observe that there can be many choices for @xmath646 ) .
it follows that @xmath647 , and this gives the required isomorphism .
[ ifcirclingaroundtrivialcohomology ] let @xmath55 be a connected complex of polytopes obtained by gluing finitely many polytopes @xmath648 of the same dimension .
assume that all these polytopes share a fixed face of codimension @xmath132 and that they are organized around it as follows : for @xmath649 , @xmath51 and @xmath650 share a facet , and also @xmath651 and @xmath652 do .
then @xmath280 .
let @xmath653 be a cocycle and consider a coboundary with @xmath654 and the other @xmath655 equal to @xmath54 . then @xmath656 .
we can iterate this strategy until we obtain @xmath657 for some @xmath658 .
now consider a coboundary with @xmath659 , @xmath660 and the other @xmath655 equal to @xmath54 .
this makes sense because @xmath661 by the cocycle condition on @xmath662 .
we can conclude that @xmath663 .
let @xmath55 be any polyhedral subdivision of @xmath31 where @xmath21 is the unit cube . applying lemma [ removehangingpolytope ] iteratively
, we can reduce ourself to compute @xmath664 with @xmath640 without hanging polytopes ( see definition [ definitionhangingpolytope ] ) . now , after enumerating all the possibilities for @xmath55 , one can observe that either @xmath640 is generated by a single maximal dimensional polytope , or it is as in the statement of lemma [ ifcirclingaroundtrivialcohomology ] . in both cases , we can conclude that @xmath665 . a complete list of all the possibilities for @xmath55 can be recovered from the table in figure [ topologicaltypesubdivisions ] .
a polyhedral subdivision of @xmath31 is represented by a graph whose vertices correspond to the maximal dimensional polytopes in the subdivision and there is an edge between two vertices if and only if the corresponding polytopes share a facet .
when this graph is not enough to distinguish two distinct polyhedral subdivision , we associate a number to each vertex which represent the number of facets of the corresponding polytope .
if this is not enough , instead we draw a dotted edge between two vertices if the corresponding polytopes share an edge and not a facet .
the sum in the top left corner of each box in figure [ topologicaltypesubdivisions ] represents the lattice volume of the polytopes used in the subdivision .
if @xmath395 is singular , assume without loss of generality that there is a singular point in the affine patch where @xmath672 and @xmath673 .
let @xmath674 , @xmath675 , and @xmath676 be the coordinates in this affine patch . then the equation of @xmath395 becomes @xmath677 therefore the following system of equations has a solution : @xmath678 this implies that @xmath679 , and therefore the two vectors @xmath680 and @xmath681 are proportional because the matrix @xmath682 has rank @xmath54 . if @xmath683 for @xmath684 , then the equation of @xmath395 becomes @xmath685 where @xmath686 and @xmath687 are glued along a ruling of @xmath460 and the line @xmath688 in @xmath2 .
the restrictions of @xmath434 and @xmath435 to these two irreducible components are described in remark [ explicitdescriptionirreduciblecomponentsdegenerations ] ( see also figure [ explicitdegenerationsinvariantwallcrossing ] ) . in this case , to conclude that @xmath409 is stable , we reduce the question to each irreducible component of @xmath395 and we apply what we already proved in the @xmath410 and @xmath428 cases ( see the proofs of propositions [ adescription ] and [ bdescription ] ) .
now let us assume that @xmath395 is smooth ( and hence irreducible ) . by the discussion above , the two vectors @xmath680 and @xmath681 are not proportional .
denote by @xmath318 the point of intersection of the two lines @xmath688 and @xmath689 in @xmath2 .
if @xmath690 is the restriction to @xmath395 of the usual projection map @xmath691 , observe that @xmath86 is an isomorphism if restricted to the complement of @xmath692 and @xmath693 .
this proves that @xmath694 . in this case , let us explain how @xmath461 depends on the coefficients @xmath695 .
the restriction @xmath461 has equation @xmath696 by using an argument completely analogous to what we just did for @xmath395 , we have that this restriction is irreducible if and only if @xmath697 and @xmath698 are not proportional . in this case , @xmath461 is a section of @xmath395 with self - intersection @xmath54 .
if @xmath697 and @xmath698 are proportional , then the equation of @xmath461 becomes @xmath699 for some @xmath684 .
the irreducible component @xmath700 ( resp .
@xmath701 ) is the exceptional section ( resp . a fiber ) of @xmath395 .
we conclude that @xmath409 is log canonical .
now we consider the case where exactly one among @xmath669 and @xmath670 is zero .
it is easy to check that @xmath395 is automatically smooth and a similar description to the one above applies , with the only difference that the restriction @xmath462 can acquire a triple intersection point .
but then we can apply lemma [ threeincidentlinesweightslogcanonical ] to conclude that @xmath409 is log canonical .
consider the blow up morphism @xmath706 and let @xmath425 be the exceptional divisor .
if @xmath707 , then we have that @xmath708 this implies that @xmath705 is log canonical because @xmath104 is a log resolution of singularities .
@xmath709,[y_0:y_1],[z_0:z_1]),\\ d_p^\bullet&=\emptyset~\textrm{(there is no conductor divisor in this case)},\\ \delta_p^\bullet&=v(x_0x_1y_0y_1z_0z_1),\\ b_p^\bullet&=v\left(\sum_{i , j , k=0,1}c_{ijk}x_iy_jz_k\right),\end{aligned}\ ] ] where any two coefficients @xmath253 and @xmath710 can not be simultaneously zero if @xmath711 and @xmath712 are vertices of the same edge of the cube ( this is because inside @xmath21 we can not fit two corner cuts with apices lying on the same edge ) .
let us first assume that @xmath395 is smooth ( hence irreducible ) , which happens for a generic choice of the coefficients @xmath253 by bertini s theorem ( see ( * ? ? ?
* chapter ii , theorem 8.18 ) ) .
then the anticanonical class @xmath713 is ample and @xmath714 , implying that @xmath715 ( see ( * ? ? ?
* exercise v.21(1 ) ) ) .
if all the @xmath253 are nonzero , then the restriction @xmath716 can be as in proposition [ mainconstruction ] ( generic case ) , or some of these lines can break into the union of two incident @xmath178-curves .
if some coefficients @xmath253 are zero , then the lines configuration @xmath716 acquires triple intersection points . in any case
, the pair @xmath717 is log canonical ( here we use lemma [ threeincidentlinesweightslogcanonical ] in case of triple intersection points ) .
now assume that @xmath395 is irreducible but singular .
let @xmath718 be a singular point . we prove that @xmath318 is a singularity of type @xmath472 and the only singular point of @xmath395 .
up to obvious symmetries , we can assume that @xmath318 is in the form @xmath719,[1:b],[1:c])$ ] .
consider the invertible change of coordinates @xmath720 , @xmath721 and so on .
this automorphism sends @xmath395 to an isomorphic surface @xmath722 which is singular at @xmath723,[1:0],[1:0])$ ] .
if we set @xmath724 and @xmath725 , then the equation of @xmath722 in this affine patch is in the form @xmath726 the coefficients @xmath727 and @xmath728 are nonzero because @xmath722 is irreducible .
therefore it is clear that the singularity is of type @xmath472 .
now we show that @xmath729 is the only singularity of @xmath722 , whose equation is given by @xmath730 where we intentionally dropped the prime " signs for simplicity of notation .
it is enough to check the affine patches @xmath731 and @xmath732 because of the symmetries of our equation . * in @xmath733
we have @xmath734 let us show that this system has no solution other than @xmath735 .
observe that @xmath736 or @xmath737 implies that @xmath738 .
therefore , assume @xmath739 and @xmath740 nonzero .
but if we multiply the last equation of the system by @xmath740 , then the first equation implies that @xmath741 , which is a contradiction .
* in @xmath742 we have to study the solutions of @xmath743 but this system has no solutions because @xmath744 or @xmath745 implies @xmath746 or @xmath747 respectively . * in @xmath748 consider @xmath749 which has no solutions because @xmath744 implies @xmath746 . * in the affine patch @xmath732 , we obviously have no singular points because the dehomogeneization of equation ( [ homogeneized ] ) is linear .
it is easy to see that the singular point @xmath318 can lie on at most one irreducible component of @xmath716 . to prove that
@xmath717 is log canonical , we show that @xmath750 is log canonical and then we use inversion of adjunction ( see @xcite ) .
this is done in two steps .
* @xmath750 is log canonical at a quadruple intersection point @xmath318 of @xmath751 .
assume without loss of generality that @xmath752,[1:0],[1:0])$ ] , an therefore @xmath753 .
in an affine neighborhood of @xmath318 the equation of @xmath395 becomes @xmath754 where we must have @xmath755 and @xmath756 nonzero .
the affine equations for @xmath416 at @xmath318 are @xmath736 and @xmath737 .
therefore , locally at @xmath318 , the four irreducible components of @xmath751 are in general linear position .
now we use lemma [ fourgeneralplanesrightweights ] ; * @xmath757 is log canonical .
this is true in the complement of @xmath416 because @xmath395 has at most an @xmath472 singularity .
let @xmath758 be an irreducible component . we need to show that @xmath757 is log canonical in a neighborhood of @xmath143 .
but this follows from inversion of adjunction because @xmath759 is log canonical ( in case @xmath751 has a quadruple intersection point , we use the previous step ) .
we are left with the case @xmath395 reducible .
decompositions into two irreducible components are given by @xmath760 for any choice of nonzero coefficients such that @xmath761 .
it is easy to see that these two components are isomorphic to @xmath762 and how they are glued together .
decompositions into three irreducible components are given by @xmath763 for any choice of nonzero coefficients .
up to @xmath764 there is only one choice of coefficients . in both cases @xmath409 is semi log canonical .
9 alexeev , v. : _ boundedness and k2 for log surfaces _ , int . j. math . 5 ( 1994 )
6 . : _ log canonical singularities and complete moduli of stable pairs _ ( 1996 ) .
arxiv : alg - geom/9608013 : _ moduli spaces @xmath768 for surfaces_. higher - dimensional complex varieties ( trento , 1994 ) , de gruyter , berlin , 1996 , pp .
complete moduli in the presence of semiabelian group action_. ann . of math .
( 2 ) 155 ( 3 ) ( 2002 ) , 611708 . :
_ higher - dimensional analogues of stable curves_. international congress of mathematicians . vol .
ii , 515536 , eur .
zrich , 2006 . : _ weighted grassmannians and stable hyperplane arrangements _ , arxiv:0806.0881 . : _ moduli of weighted hyperplane arrangements _ , springer - birkhuser , 2015 .
alexeev , v. , pardini , r. : _ explicit compactifications of moduli spaces of campedelli and burniat surfaces_. ( 2009 ) .
arxiv:0901.4431 : _ non - normal abelian covers_. compos .
4 , 10511084 ( 2012 ) alexeev , v. , thompson , a. : _ modular compactification of moduli of k3 surfaces of degree 2 _ , manuscript in preparation ( 2015 ) . barth , w.p .
, hulek , k. , peters , c.a.m .
, van de ven , a. : _ compact complex surfaces_. second edition .
ergebnisse der mathematik und ihrer grenzgebiete .
3 . folge . a series of modern surveys in mathematics [ results in mathematics and related areas .
3rd series .
a series of modern surveys in mathematics ] , 4 .
springer - verlag , berlin , 2004 .
beauville , a. : _ complex algebraic surfaces_. translated from the 1978 french original by r. barlow , with assistance from n. i. shepherd - barron and m. reid .
second edition .
london mathematical society student texts , 34 . cambridge university press , cambridge , 1996 .
cossec , f.r .
: _ on the picard group of enriques surfaces_. math .
( 1985 ) , no .
4 , 577600 .
de loera , j.a . : _
nonregular triangulations of products of simplices_. discrete comput .
( 1996 ) , no .
3 , 253264 .
de loera , j.a . ,
rambau , j. , santos , f. : _ triangulations : structures for algorithms and applications_. algorithms and computation in mathematics , 25 , springer - verlag , berlin ( 2010 ) .
dolgachev , i. : _ classical algebraic geometry . a modern view_. cambridge university press , cambridge , 2012 .
ferrand , d. : _ conducteur , descente et pincement_. bull .
france 131 ( 2003 ) , no .
4 , 553585 . friedman , r. : _ base change , automorphisms , and stable reduction for type iii k3 surfaces_. the birational geometry of degenerations ( cambridge , mass . , 1981 ) ,
277298 , progr .
, 29 , birkhuser , boston , mass . ,
friedman , r. , morrison , d.r .
: _ the birational geometry of degenerations : an overview_. the birational geometry of degenerations ( cambridge , mass . , 1981 ) , pp .
132 , progr .
, 29 , birkhuser , boston , mass . ,
fujino , o. : _ semipositivity theorems for moduli problems _ ( 2015 ) .
giansiracusa , n. , gillam , w.d .
: _ on kapranov s description of @xmath769 as a chow quotient_. turkish j. math .
38 ( 2014 ) , no .
4 , 625648 .
kapranov , m.m . ,
strumfels , b. , zelevinsky , a.v . : _
discriminants , resultants and multidimensional determinants_. reprint of the 1994 edition .
modern birkhuser classics .
birkhuser boston , inc . ,
boston , ma , 2008 . hacking , p. : _ compact moduli of plane curves_. duke math .
j. 124 ( 2004 ) , no .
2 , 213257 . hacking , p. , keel , s. , tevelev , j. : _ stable pair , tropical , and log canonical compactifications of moduli spaces of del pezzo surfaces_. invent .
178 ( 2009 ) , no .
1 , 173227 .
hacon , c. , mckernan , j. , xu , c. : _ boundedness of moduli of varieties of general type _ ( 2014 ) .
arxiv:1412.1186 hartshorne , r. : _ algebraic geometry_. springer - verlag , new york , 1977 , graduate texts in mathematics , 52 .
hassett , b. : _ moduli spaces of weighted pointed stable curves_. adv . math . 173
( 2003 ) , no .
2 , 316352 . hu ,
x. : _ the compactifications of moduli spaces of burniat surfaces with @xmath770 _ ( 2014 ) .
kawakita , m. : _ inversion of adjunction on log canonicity_. invent .
( 2007 ) , no .
1 , 129133 .
kollr , j. : _ hulls and husks _ ( 2009 ) .
arxiv:0805.0576 : _ moduli of varieties of general type_. handbook of moduli . vol .
ii , 131157 , adv .
lect . math .
( alm ) , 25 , int . press , somerville ,
ma , 2013 . : _ singularities of the minimal model program_. with a collaboration of sndor kovcs .
cambridge tracts in mathematics , 200 .
cambridge university press , cambridge , 2013 .
kollr , j. , mori , s. : _ birational geometry of algebraic varieties _ ,
134 of cambridge tracts in mathematics , cambridge university press , 2008 .
kollr , j. , shepherd - barron , n. : _ threefolds and deformations of surface singularities _ , invent .
91 ( 1988 ) , no .
2 , 299338 .
kovcs , s. , patakfalvi , z. : _ projectivity of the moduli space of stable log - varieties and subadditvity of log - kodaira dimension _ ( 2015 ) .
kulikov , v.s . : _ degenerations of k3 surfaces and enriques surfaces_. izv .
nauk sssr ser . mat .
41 ( 1977 ) , no . 5 , 10081042 , 1199 .
laza , r.:_the ksba compactification for the moduli space of degree two k3 pairs_. j. eur .
( jems ) 18 ( 2016 ) , no .
2 , 225279 .
lazarsfeld , r. : _ positivity in algebraic geometry i _ , classical setting : line bundles and linear series , ergebnisse der mathematik und ihrer grenzgebiete .
3 . folge . a series of modern surveys in mathematics [ results in mathematics and related areas .
3rd series .
a series of modern surveys in mathematics ] , 48 , springer - verlag , berlin ( 2004 ) looijenga , e. : _ compactifications defined by arrangements .
locally symmetric varieties of type iv_. duke math .
j. 119 ( 2003 ) , no .
3 , 527588 . matsumoto , k. , sasaki , t. , yoshida , m. : _ the monodromy of the period map of a @xmath0-parameter family of k3 surfaces and the hypergeometric function of type @xmath771_. internat . j. math . 3 ( 1992 ) , no . 1 , 164 .
mumford , d. : _ the red book of varieties and schemes_. second , expanded edition . includes the michigan lectures ( 1974 ) on curves and their jacobians . with contributions by enrico arbarello .
lecture notes in mathematics , 1358 .
springer - verlag , berlin , 1999 .
oudompheng , r. : _ periods of an arrangement of six lines and campedelli surfaces _ , universit de nice - sophia antipolis , phd thesis , part ii ( 2010 ) .
pardini , r. : _ abelian covers of algebraic varieties_. j. reine angew .
( 1991 ) , 191213 .
persson , u. , pinkham , h. : _ degeneration of surfaces with trivial canonical bundle_. ann . of math .
( 2 ) 113 ( 1981 ) , no .
1 , 4566 .
pfeifle , j. : _ secondary polytope of the 3-cube _ , electronic geometry model no .
2000.09.031 . `
http://www.eg-models.de/models/polytopes/secondary_polytopes/2000.09.031/_preview.html ` santos , f. : _ on the refinements of a polyhedral subdivision_. collect .
52 ( 2001 ) , no .
3 , 231256 .
scattone , f. : _ on the compactification of moduli spaces for algebraic k3 surfaces_. mem .
70 ( 1987 ) , no .
374 , x+86 pp .
shah , j. : _ a complete moduli space for k3 surfaces of degree 2_. ann . of math .
( 2 ) 112 ( 1980 ) , no .
3 , 485510 . shah , j. : _ degenerations of k3 surfaces of degree 4_. trans .
( 1981 ) , no .
2 , 271308 .
shepherd - barron , n.i . : _ extending polarizations on families of k3 surfaces_. the birational geometry of degenerations ( cambridge , mass . , 1981 ) , pp . 135171 , progr .
, 29 , birkhuser , boston , mass . , 1983 .
viehweg , e. : _ quasi - projective moduli for polarized manifolds_. vinberg , .b . : _ some arithmetical discrete groups in lobaevskispaces_. discrete subgroups of lie groups and applications to moduli ( internat .
bombay , 1973 ) , pp .
oxford univ . press , bombay , 1975 . | we describe the moduli compactification by stable pairs of a @xmath0-dimensional family of enriques surfaces , which arise as the @xmath1-cover of the blow up of @xmath2 at three general points branched along a configuration of six lines .
the divisor is chosen to be an appropriate multiple of the ramification locus .
we study the degenerations parametrized by the boundary and its stratification , which has a combinatorial interpretation in terms of the polyhedral subdivisions of the unit cube .
finally , we relate this compactification to the baily - borel compactification of the same @xmath0-dimensional family of enriques surfaces . |
large - scale magnetic fields with micro gauss strength have been observed in galaxies @xcite and clusters of galaxies @xcite .
although it is assumed that such magnetic fields are the result of amplification from weak seed magnetic fields , the origin of the seed magnetic field has not yet been understood .
there are two classes of models for the seed generation mechanism : astrophysical and cosmological models . in astrophysical models ,
the seeds are associated with nonlinear structures , and are generated during structure formation through the biermann battery effect @xcite , or the weibel instability @xcite . on the other hand ,
seeds in the cosmological models are produced in the early universe and exist as extragalactic magnetic fields .
this class of models includes the generation mechanisms in an inflationary epoch , during cosmological phase transitions , and at the epoch of recombination ( for recent reviews , see @xcite ) .
recent gamma ray observations suggest the existence of cosmological magnetic fields stronger than @xmath0 gauss in the voids @xcite .
although more detailed work will be required to establish this lower limit @xcite , the suggestion of such magnetic fields in large - scale structure voids , @xmath1 away from non - linear structures , provides a strong argument in favor of cosmological models . further support in favor of an early universe origin can be obtained if the cosmological galactic magnetic fields are `` helical '' , thus indicating a process of magneto - genesis that fundamentally violates invariance under charge conjugation plus parity reflection ( cp ) @xcite .
magnetic helicity density is defined as @xmath2 where @xmath3 is the vector potential of magnetic fields @xmath4 with @xmath5 .
magnetic helicity is odd under cp transformations as @xmath6 and @xmath4 are odd under c , while @xmath6 is even but @xmath4 is odd under p. non - zero magnetic helicity is predicted in scenarios in which cosmic baryogenesis and magneto - genesis occur concurrently during a cosmological phase transition @xcite . then the magnetic helicity density is related to the cosmic baryon number density and the cp violation responsible for the excess of matter over antimatter also provides helicity to the magnetic field .
other scenarios that can generate helical magnetic fields that are not tied to baryogenesis have been studied in refs .
helicity is also an important factor in the evolution of magnetic fields because it helps to transfer magnetic field energy from small to large length scales , a process called an `` inverse cascade '' . due to the inverse cascade
, helical magnetic fields can grow to astrophysically relevant scales at the present epoch , even though the initial scale of the magnetic field extends only up to the much smaller cosmological horizon scale at the time of the phase transition .
the two point correlator of a stochastic , homogeneous , isotropic magnetic field contains two independent functions . in physical space , these are the normal and helical correlation functions ( @xmath7 and @xmath8 ) ; in momentum space , these are the symmetric and antisymmetric power spectra ( @xmath9 and @xmath10 ) .
there are several tools to measure the normal correlator ( or the symmetric power spectrum ) of cosmological magnetic fields but measuring helicity directly is more challenging .
the detection of helicity of cosmological magnetic fields has been a primary motivation for this work , as few other direct schemes to detect helicity have been proposed @xcite .
a measurement of the normal correlator itself , together with some theoretical input , provides indirect access to the helicity of the magnetic field .
this is because the presence of helicity influences the evolution of the normal correlator , and the exponents characterizing the normal correlator can give us some information about the magnetic helicity ( for the evolution in the cosmological context , see ref.@xcite ) .
the shape of the normal correlator , in particular the existence of a peak in the distribution can also inform us about the epoch of magneto - genesis . if magnetic fields are produced during inflation , the symmetric power spectrum can be expected to be scale invariant with significant power on the present day horizon scale .
direct measures of the helical power spectrum are more difficult . for _ astrophysical _ magnetic helicity , it has been suggested that the correlation between faraday rotation measurement and the polarization degree of radio synchrotron emission @xcite can be used . for _
cosmological _ magnetic helicity , ref .
@xcite has shown that correlations in the arrival momenta of very high energy cosmic rays are sensitive to the intervening magnetic helicity provided the cosmic ray source locations are known .
cosmic microwave background ( cmb ) anisotropy observations may also permit a measure of the helicity through non - vanishing cross - correlation between the temperature and b - mode polarization anisotropies and between the e - mode and b - mode polarization anisotropies @xcite . in the present paper ,
we propose a scheme to measure both the normal and the helical correlation functions of cosmological magnetic fields by using tev blazar observations . gamma rays with energy greater than @xmath11
can scatter with ambient `` extragalactic background light '' ( ebl ) photons to pair produce electrons and positrons .
the generated electrons and positrons create a secondary cascade of gev gamma rays through the inverse compton ( ic ) scattering of cmb photons . as the electron and positron trajectories
are bent due to the lorentz force by a magnetic field , the gev photon cascade carries information about the structure of the cosmological magnetic field . we show that cross - correlations between the arrival directions of the secondary cascade gamma rays at different energies are related to the correlation function of the extragalactic magnetic fields .
if we imagine the cascade photon arrival direction to be a vector in the plane of observation , the inner product of vectors at different energies gives the normal part of the magnetic field correlation , and the outer product gives the helical correlator . in sec .
[ geometry ] we describe the basic geometry of the process and evaluate the arrival direction of cascade photons as a function of the intervening magnetic field . in sec .
[ correlators ] we evaluate correlators of the arrival directions of cascade photons and relate them to the magnetic field correlation functions .
our results are placed in the context of observations in sec .
[ observations ] where we discuss estimators for the theoretical correlation functions we have found in sec .
[ correlators ] .
our analysis uses many simplifying assumptions that we discuss together with conclusions in sec .
[ conclusions ] . throughout this paper , we use natural units : @xmath12 .
tev gamma rays from distant sources at redshift @xmath13 can not propagate freely over cosmological distances because , at such energies , interaction with the ebl can produce electrons and positrons .
the mean free path of a gamma ray with energy @xmath14 is given by @xcite @xmath15 where @xmath16 is a numerical factor which accounts for the model uncertainties of ebl .
here we take @xmath17 @xcite .
electrons and positrons generated by the tev gamma ray lose energy by the production of a secondary gamma ray cascade through the ic scattering of cmb photons , and have a mean free path , @xmath18 where @xmath19 is the typical redshift at which tev gamma rays create pairs , and @xmath20 is the electron energy .
the up - scattered cmb photon has energy , @xmath21 where @xmath22 is the typical energy of cmb photons .
therefore we observe cascade gamma rays with energy @xmath23 , in addition to tev gamma rays with energy @xmath24 .
it is believed that tev gamma rays are beamed from distant blazars in a narrow jet of opening angle @xmath25^{-1}$ ] where @xmath26 is the lorentz factor of the gamma ray emitting plasma .
the highest energy photons are observed if we lie within the opening angle of the jet . denoting the angle between the source direction and the jet axis by @xmath27
, we require @xmath28 . as emphasized in ref .
@xcite , the most likely situation is that we are located on the edge of the cone , as depicted in fig .
[ fig : tri ] . .
the observer is most likely located at the edge of the jet , not on the axis .
tev photons pair produce after propagating a distance @xmath29 .
the pairs are bent by ambient magnetic fields and up - scatter cmb photons that propagate a distance @xmath30 to the observer .
the emission angle @xmath31 , the observation direction @xmath32 , the distance to the source @xmath33 , and the pair creation and ic scattering event positions , @xmath34 , @xmath35 are also shown.,width=377 ] suppose that the blazar is located at @xmath36 in the observer frame .
we will assume for simplicity that the redshift of the source is less than one and take @xmath37 .
if there are no magnetic fields in the igm , we will observe cascade gamma rays due to tev gamma rays also in the direction @xmath38 .
however if cosmological magnetic fields are present , the cascade gamma rays arrive from different directions .
we now evaluate the arrival direction assuming a stochastic , homogeneous , and isotropic magnetic field .
consider an observed cascade gamma ray that resulted from a tev gamma ray with energy @xmath14 that was emitted at @xmath39 at an angle @xmath31 from the line of sight as depicted in fig .
[ fig : tri ] .
the tev gamma ray produces an electron at the position @xmath40 at time @xmath41 , where @xmath42 .
the produced electron has momentum @xmath43 whose amplitude @xmath44 corresponds to the energy @xmath45 . since the opening angle of the electron - positron pair is very small , of order @xmath46 , the direction @xmath47 is the same as the direction of the initial tev gamma ray . the momentum of the electron changes on propagation due to the lorentz force , @xmath48 where @xmath49 is the electron / positron charge , @xmath50 and @xmath51 are the position and velocity of the electron ( or positron ) , namely @xmath52 where the overdot denotes differentiation with respect to time .
( for convenience , from now on we shall refer to the charged particle as being the electron . )
we will now decompose all vectors in components parallel and perpendicular to the source direction .
for example , the momentum and the position of the electron at time @xmath53 is decomposed as @xmath54 where the subscript @xmath55 means the component parallel to the source direction ( line - of - sight for tev source ) .
therefore , the vector @xmath56 and @xmath57 are the deviations induced by the magnetic field . in terms of the decomposed components , eq .
( [ eq : momentum_evo ] ) can be written as @xmath58 \times { \bm b}({\bm x}(t ' ) ) , \label{eq : momentum_evo_m}\ ] ] where @xmath59 and @xmath60 are the momentum components at time @xmath41 .
note that at this stage , instead of replacing @xmath61 by @xmath62 in the argument of @xmath4 , we perform the integration along the actual path @xmath63 .
this is important if the magnetic fields have significant power on small scales , _
i.e. _ , a blue spectrum .
the bending angle of the electron is estimated as @xmath64 [ e_{\rm tev}/10~{\rm tev}]^{-2}$ ] where @xmath65 is the larmor radius .
here we have assumed a magnetic field coherence scale larger than @xmath66 ; otherwise the electron trajectory would be diffusive yielding a smaller estimate for @xmath67 .
then the maximum deviation from the source direction is @xmath68 [ e_{\rm tev}/10~{\rm tev}]^{-3}$ ] .
since the bending angle is small , we can treat @xmath69 , @xmath70 and @xmath71 as perturbations .
to linear order in the magnetic field strength , eq .
( [ eq : momentum_evo_m ] ) becomes @xmath72 the electron energy @xmath73 is constant during this process , since a magnetic field does no work .
dividing eq .
( [ eq : d_momentum_evo ] ) by @xmath73 , we obtain the velocity , @xmath74 and another integration gives the trajectory , @xmath75 at any point in the electron s trajectory there is a probability of an ic scattering event with a cmb photon .
we simplify the present analysis by assuming that the electron travels a fixed distance @xmath76 , and at final time @xmath77 up - scatters a cmb photon .
since the electron is ultra - relativistic with lorentz boost factor @xmath78 , the cascade gamma ray propagates along the momentum of the electron . therefore the arrival direction of the cascade gamma ray is the same as the direction of the position @xmath79 .
we now define the vector @xmath80 in the observation plane by @xmath81 where @xmath82 is @xmath83 at @xmath84 .
the magnitude , @xmath85 , is the observed angle @xmath32 shown in fig .
[ fig : tri ] , and the direction in the observation plane corresponds to the azimuthal direction of the original tev gamma ray
. we will now relate @xmath80 to the magnetic field using eq .
( [ eq : position_evo_m ] ) .
however @xmath86 is still unknown in eq .
( [ eq : position_evo_m ] ) . to evaluate it
, we first need to express the emission angle @xmath31 in terms of the observed angle @xmath32 . applying the trigonometric law of sines to the triangle formed by the source , observer and electron position
recall that @xmath87 we obtain @xmath88 in terms of @xmath80 , the vector @xmath89 can now be written as @xmath90 where @xmath91 is the magnitude of the electron velocity , and @xmath92 is the unit vector perpendicular to the line - of - sight ( see fig . [
fig : tri ] ) . now eqs .
( [ eq : position_evo_m ] ) , ( [ eq : theta_obs ] ) and ( [ eq : vi ] ) give , @xmath93 where we can rewrite @xmath29 , @xmath66 , and @xmath73 as functions of @xmath94 , @xmath95 @xmath96 @xmath97 with these relations , the magnitude of @xmath98 is roughly estimated as @xmath99
the vector @xmath100 describes the position of the observed cascade gamma ray in the observation plane but its magnitude depends on geometrical factors such as the distance to the source ( eq . ( [ eq : final_ang ] ) ) .
we define a rescaled vector to remove such dependence , @xmath101 now we are interested in two types of correlators of the @xmath102 vectors @xmath103 @xmath104 where @xmath105 and @xmath106 are two energies of the observed cascade gamma rays , and we recall that we have set up our coordinate system so that the @xmath107axis is along the line of sight , _
i.e. _ , @xmath108 is normal to the observation plane .
the ensemble value @xmath109 , where @xmath110 and @xmath111 denote components in the observation plane , can be written by using the magnetic field correlation function , eq .
( [ eq : final_ang ] ) , @xmath112 where @xmath113 and @xmath114 are @xmath66 for the cascade gamma ray with energy @xmath105 and @xmath106 respectively .
note that the velocity @xmath115 and @xmath116 .
the correlation function of a stochastic , homogeneous , and isotropic magnetic field is given by @xcite @xmath117 + m_l(r ) { r_i r_j \over r^2 } + m_h ( r ) \epsilon_{ijl } r^l , \label{bbcorr}\ ] ] where @xmath7 , @xmath118 , and @xmath8 are the correlation functions for the normal , longitudinal , and helical parts of the magnetic fields .
due to the homogeneity and isotropy of the magnetic fields , these correlations depend only on the separation distance @xmath119 .
the divergence - less condition gives @xmath120 the stochastic fields are often described by power spectra in fourier space . the magnetic field correlation function in fourier space is @xmath121,\ ] ] where @xmath9 and @xmath10 are the symmetric and antisymmetric ( helical ) parts of the magnetic field power spectrum .
the functions @xmath9 and @xmath10 are related to the correlation functions @xmath7 ,
@xmath118 , and @xmath8 as in @xcite . therefore the ensemble average , eq .
( [ eq : ensemble ] ) , can also be decomposed into three parts : the normal , longitudinal , and helical parts , @xmath122 with each of the terms given by four integrations as in eq .
( [ eq : ensemble ] ) .
the separation scale , @xmath123 , appearing in the correlators @xmath7 , @xmath118 and @xmath8 is the magnitude of the separation vector @xmath124 since cosmological magnetic fields are weak , this can be approximated as @xmath125 therefore , @xmath126 for @xmath127 and @xmath128 , the separation scale becomes @xmath129 in the case with @xmath130 , @xmath131 and @xmath132 vanish , because only the @xmath108 components of @xmath133 and @xmath134 are non - vanishing .
however the normal correlator does not vanish and @xmath135 is given by @xmath136 in the case with @xmath137 , @xmath138 and @xmath139 , while @xmath140 and @xmath141 vanish , @xmath142 has non - zero value , @xmath143 therefore @xmath144 and @xmath145 can be written directly in terms of the magnetic field correlation functions , @xmath146 @xmath147 it is worth pointing out that @xmath148 but @xmath149 .
we have already seen that @xmath150 independently of @xmath94 , and so the separation scale in the magnetic field correlation function can be approximated as , @xmath151 then recalling @xmath152 , we can do the integrations in eqs .
( [ eq : f ] ) and ( [ eq : g ] ) to get @xmath153 @xmath154 where @xmath155 is @xmath156 .
this shows that we can obtain both the non - helical and the helical magnetic field correlation functions through the correlators @xmath157 and @xmath158 constructed from the arrival information of cascade photons .
this is the main result of this paper .
note , as shown in fig .
[ fig : cor ] , that the obtained correlator is on the scale @xmath155 which can be much larger than the scale @xmath66 .
this is because the magnetic field gets correlated at the spatial points where pair production occurs and these can be separated by hundreds of mpc depending on the values chosen for @xmath105 and @xmath106 .
it remains an interesting open question if further refinements of the above method can lead to correlators on scales smaller than @xmath159 .
( solid lines at the vertices of the triangles ) .
the regions themselves are separated by distance @xmath123 which can be @xmath1 depending on the energy difference of the two events .
energy resolution of the detector translates into a lower limit on the separation at which the magnetic field correlations can be probed .
, width=377 ]
in the previous section , we have considered correlators in the arrival directions of cascade photons and expressed them in terms of correlators of the intervening magnetic field . in this section
we consider the problem of determining magnetic field correlators from the observers point of view : given cascade arrival direction data , what quantity should be calculated that corresponds to the magnetic field correlator ?
the calculated quantity will at best be an estimator for the magnetic field correlator because observations are limited in number , while the magnetic field correlator in eq .
( [ bbcorr ] ) is over an infinite ensemble of realizations . in a more realistic setting
see simulations in ref .
@xcite observed gamma rays with some energy @xmath160 are expected to be scattered around a typical observation direction on the observational plane .
the scattering is due to differences in the magnetic field along the trajectories of different electrons and also the stochasticity of the ebl and cmb photons .
we have not considered the latter fluctuations in this paper but the variation in the magnetic field can be smoothed out by taking the average @xmath161 over all observed gamma rays with the same energy @xmath160 . even if the magnetic fields have small - scale structure , _
i.e. _ , a blue spectrum , the averaging procedure should yield the correct magnetic field correlator on scales larger than the smoothing scale , @xmath162 .
there is a second way to smooth out fluctuations in the magnetic field .
this is by realizing that the correlation functions @xmath163 and @xmath164 evaluated in the previous section depend on only one function , @xmath165 .
hence there are many choices of @xmath105 and @xmath106 yielding the same @xmath123 , and one of the energy variables can be integrated out .
this corresponds to averaging over all pairs of energies such that the distance @xmath123 remains fixed ( see fig . [
fig : cor ] ) . using eq .
( [ re1e2 ] ) we see that @xmath166 is independent of @xmath94 provided @xmath167.\ ] ] we can now average eqs .
( [ eq : f-2 ] ) and ( [ eq : g-2 ] ) over @xmath105 while taking @xmath168 , @xmath169 @xmath170 where @xmath171 is the integration range of the observation energy .
the vector @xmath172 now denotes the _ average _ ( rescaled ) direction vector for cascade photons with energy @xmath160 . a third way to perform
the ensemble average is to use observations of many tev blazars because cascade gamma rays from different blazars sample magnetic fields along different path .
hence we can obtain the magnetic field correlation function using @xmath173_\alpha , \label{eq : observe_ave1}\ ] ] @xmath174_\alpha , \label{eq : observe_ave}\ ] ] where @xmath175 labels the blazar .
finally we would like to remark that the most likely observation of a tev blazar is when we are positioned at the edge of the jet @xcite , but we are even more likely to be located outside the jet opening angle
. then we will not observe the tev source but will still receive some cascade gev photons . in this case , if there is reason to suppose that the observed photons are indeed from a cascade , we can extend our correlator by replacing @xmath172 in eqs .
( [ eq : observe_ave1 ] ) , ( [ eq : observe_ave ] ) by @xmath176 where @xmath177 is the highest energy ( and least deviated ) cascade photon that is observed
. we can still draw some conclusions about the magnetic field correlators though the analysis is more involved .
our main result is the connection between cosmological magnetic field correlators and the correlators of cascade photons as given in eqs .
( [ eq : f-2 ] ) and ( [ eq : g-2 ] ) together with the definitions in eqs .
( [ fdef ] ) and ( [ gdef ] ) . in the observational context
, this connection can be written as in eqs .
( [ eq : observe_ave1 ] ) and ( [ eq : observe_ave ] ) .
these relations suggest that observations of cascade photons may be used to directly study the non - helical and helical spectra of cosmological magnetic fields .
there are few other direct ways to probe the helicity of cosmological magnetic fields and this is an important feature of the technique we have described . our analysis will need further development as it has used many simplifying assumptions .
for example , the stochastic interactions of the tev photons with the ebl photons will lead to a probability distribution for @xmath29 while we have used a fixed value .
similarly there will be probability distributions for the other interactions , and the magnetic field will have some structure on small scales .
in addition , if the blazar is at high redshift , cosmological expansion will become important .
these effects can best be studied using monte carlo simulations similar to the ones described in refs .
the simulations should be able to confirm if the magnetic field correlators can be recovered from realistic data , which hopefully will become available in the near future .
m. l. bernet , f. miniati , s. j. lilly , p. p. kronberg and m. dessauges - zavadsky , nature * 454 * , 302 ( 2008 ) [ arxiv:0807.3347 [ astro - ph ] ] .
t. e. clarke , p. p. kronberg and h. boehringer , astrophys .
j. * 547 * , l111 ( 2001 ) [ astro - ph/0011281 ] .
a. bonafede , l. feretti , m. murgia , f. govoni , g. giovannini , d. dallacasa , k. dolag and g. b. taylor , a&a * 513 * , a30 ( 2010 ) arxiv:1002.0594 [ astro-ph.co ] .
p. l. biermann and c. f. galea , `` the early universe and the cosmic microwave background : theory and observations '' proceedings of the nato advanced study institute , ( 2003 ) astro - ph/0302168 .
m. lazar , r. schlickeiser , r. wielebinski , s. poedts , astrophys .
j. * 693 * , 1133 , 2009 a. neronov , i. vovk and , science * 328 * , 73 ( 2010 ) [ arxiv:1006.3504 [ astro-ph.he ] ] .
f. tavecchio , g. ghisellini , l. foschini , g. bonnoli , g. ghirlanda and p. coppi , mon . not .
* 406 * , l70 ( 2010 ) [ arxiv:1004.1329 [ astro-ph.co ] ] .
k. dolag , m. kachelriess , s. ostapchenko and r. tomas , astrophys .
j. * 727 * , l4 ( 2011 ) [ arxiv:1009.1782 [ astro-ph.he ] ] .
a. e. broderick , p. chang and c. pfrommer , astrophys .
j. * 752 * , 22 ( 2012 ) [ arxiv:1106.5494 [ astro-ph.co ] ] . f. miniati and a. elyiv , arxiv:1208.1761 [ astro-ph.co ] .
r. schlickeiser r. , d. ibscher , m. supsar , astrophys . j. * 758 * , 102 ( 2012 ) .
t. vachaspati , phys .
lett . *
87 * , 251302 ( 2001 ) [ astro - ph/0101261 ] .
t. vachaspati , phys .
b * 265 * , 258 ( 1991 ) .
m. joyce and m. e. shaposhnikov , phys .
* 79 * , 1193 ( 1997 ) [ astro - ph/9703005 ] .
j. m. cornwall , phys .
d * 56 * , 6146 ( 1997 ) [ hep - th/9704022 ] .
h. tashiro , t. vachaspati and a. vilenkin , phys .
d * 86 * , 105033 ( 2012 ) [ arxiv:1206.5549 [ astro-ph.co ] ] .
g. b. field and s. m. carroll , phys .
d * 62 * , 103008 ( 2000 ) [ astro - ph/9811206 ] . l. campanelli and m. giannotti , phys .
d * 72 * , 123001 ( 2005 ) [ astro - ph/0508653 ] .
l. campanelli , int .
j. mod .
d * 18 * , 1395 ( 2009 ) [ arxiv:0805.0575 [ astro - ph ] ] .
t. kahniashvili and t. vachaspati , phys .
d * 73 * , 063507 ( 2006 ) [ astro - ph/0511373 ] .
m. christensson , m. hindmarsh and a. brandenburg , astron .
nachr . * 326 * , 393 ( 2005 ) [ astro - ph/0209119 ] .
l. campanelli , phys .
d * 70 * , 083009 ( 2004 ) [ astro - ph/0407056 ] .
r. banerjee and k. jedamzik , phys .
d * 70 * , 123003 ( 2004 ) [ astro - ph/0410032 ] .
l. campanelli , phys .
lett . * 98 * , 251302 ( 2007 ) [ arxiv:0705.2308 [ astro - ph ] ] .
a. boyarsky , j. frohlich and o. ruchayskiy , phys .
lett . * 108 * , 031301 ( 2012 ) [ arxiv:1109.3350 [ astro-ph.co ] ] .
h. junklewitz and t. a. ensslin , a&a * 530 * , a88 ( 2011 ) arxiv:1008.1243 [ astro-ph.im ] .
n. oppermann , h. junklewitz , g. robbers and t. a. ensslin , a&a , 530 , a89 ( 2011 ) .
arxiv:1008.1246 [ astro-ph.im ] .
c. caprini , r. durrer and t. kahniashvili , phys .
d * 69 * , 063006 ( 2004 ) [ astro - ph/0304556 ] .
t. kahniashvili and b. ratra , phys .
d * 71 * , 103006 ( 2005 ) [ astro - ph/0503709 ] .
k. e. kunze , phys .
d * 85 * , 083004 ( 2012 ) [ arxiv:1112.4797 [ astro-ph.co ] ] .
a. elyiv , a. neronov and d. v. semikoz , phys .
d * 80 * , 023010 ( 2009 ) [ arxiv:0903.3649 [ astro-ph.co ] ] .
k. dolag , m. kachelriess , s. ostapchenko and r. tomas , astrophys . j. * 703 * , 1078 ( 2009 ) [ arxiv:0903.2842 [ astro-ph.he ] ] . | tev blazars offer an exciting prospect for discovering cosmological magnetic fields and for probing high energy processes , including cp violation , in the early universe . we propose a method for reconstructing both the non - helical and the helical magnetic field correlators using observations of cascade photons from tev blazars . |
beginning with mcshane @xcite , sums of geometric quantities over the simple closed curves on a surface @xmath0 have been shown to exhibit remarkable properties , such as being constant over the deformation space of geometric structures on @xmath0 .
let @xmath1 denote the _
fricke space , _ consisting of isotopy classes of marked complete hyperbolic structures on @xmath0 . if @xmath2 is a closed curve on a marked hyperbolic surface @xmath3 , then the function @xmath4 assigning to @xmath3 the length of the closed geodesic on @xmath3 ( or @xmath5 if @xmath2 is homotopic to a cusp ) is a basic quantity , depending on the homotopy class of @xmath2 .
see abikoff @xcite and wolpert @xcite , 3 for background on hyperbolic riemann surfaces , where @xmath1 idenfifies with the _ teichmller space _ of @xmath0 by the uniformization theorem .
mirzakhani @xcite and tan - wong - zhang @xcite express the length of a boundary component as an infinite series of functions of the lengths of interior curves .
( compare also hu - tan - zhang @xcite and wolpert @xcite , pp .
this led mirzakhani to new calculations of the weil - petersson volume of the moduli space of curves . in another direction ,
labourie - mcshane @xcite found identities involving quantities generalizing geodesic lengths for _ hitchin representations _
of @xmath6 into @xmath7 .
for @xmath8 , these quantities are the geodesic length functions for the hilbert metric on convex @xmath9-structures on @xmath0 . in contrast , our identities are the first such identities for a _ non - reductive _ lie group . closely related to hyperbolic structures on sufraces are _ flat lorentzian structures _ on @xmath10-manifolds , that is , geometric structures modeled on minkowski space @xmath11 and groups of ( lorentzian ) isometries .
a _ margulis spacetime _
@xmath12 is a quotient @xmath13 , where @xmath14 is a discrete group of lorentzian isometries which acts properly and freely on @xmath15 .
we assume @xmath16 is not solvable , as the solvable cases are easily classified ( fried - goldman @xcite ) .
then @xmath12 is an geodesically complete flat lorentzian manifold with free fundamental group @xmath16 .
it is necessarily orientable ( @xcite ) .
the linearization @xmath17 maps @xmath16 isomorphically onto a discrete subgroup of @xmath18 ( @xcite ) .
therefore , associated to @xmath12 is a complete hyperbolic surface @xmath19 . just as essential closed curves in @xmath3 are canonically homotopic to unique closed geodesics ( having length @xmath20 ) , nonparabolic closed curves in @xmath15 are canonically homotopic to closed geodesics which are necessarily spacelike . these geodesics have a natural lorentzian length @xmath21 , called the _ margulis invariant , _ as this quantity was discovered and developed by margulis @xcite .
such a geodesic is the quotient of a line @xmath22 upon which @xmath2 acts by translation of lorentzian displacement @xmath21 .
this is similar to how @xmath23 measures the minimal displacement of the linear part of @xmath2 acting isometrically on @xmath24 .
( compare the survey articles abels @xcite , charette - drumm @xcite , drumm @xcite and guritaud @xcite for background on these geometric structures . )
this paper develops an identity for margulis spacetimes for the margulis invariants @xmath21 analogous to mcshane s identity for hyperbolic surfaces .
the lie group of interest here is the _ tangent bundle _ lie group @xmath25 ( see below)a of the isometry group @xmath26 of the hyperbolic plane @xmath24 .
the lie algebra @xmath27 of @xmath28 naturally identifies with the lorentzian vector space @xmath29 , and @xmath30 identifies with the isometric action of @xmath31 on @xmath29 .
the affine space @xmath11 identifies with the quotient @xmath32 by the zero - section .
given any lie group @xmath28 with lie algebra @xmath27 , the tangent bundle of @xmath28 enjoys a lie group structure as the semidirect product @xmath33 where @xmath34 is the adjoint representation .
then @xmath27 is an abelian normal subgroup , and the zero - section @xmath35 embeds @xmath28 as a subgroup of @xmath25 which is _ not _ normal .
this can be elegantly described in terms of base change @xmath36}$ ] , where @xmath37}$ ] is the @xmath38-algebra of _ dual numbers .
_ ( here @xmath39 is an indeterminate with @xmath40 ) .
suppose @xmath28 is the group @xmath41 of @xmath38-points of an algebraic group scheme @xmath42 .
then @xmath43})$ ] and @xmath44 ) = \big({\mathsf{hom}}(\gamma , { \mathbf{g}}\big ) ( { \mathbb{r}}[\epsilon ] ) = t { \mathsf{hom}}(\gamma , g)\ ] ] is the zariski tangent bundle of the representation variety @xmath45 . in particular homomorphisms
@xmath46 correspond to a homomorphism @xmath47 and an _ infinitesimal deformation _ of @xmath48 , that is , a tangent vector in @xmath49 .
we obtain an identity for affine deformations @xmath16 of a fuchsian group @xmath50 by differentiating the basic length identity for @xmath50 . to simplify notation , define two functions @xmath51 and @xmath52 : @xmath53 [ thm : oneend ] let @xmath19 be a complete one - ended orientable hyperbolic surface . for any affine deformation of @xmath50 ,
the margulis invariants satisfy : @xmath54 here @xmath55 denotes the set of unordered pairs of isotopy classes of simple closed curves such that @xmath56 and @xmath57 cobound an embedded @xmath10-holed sphere in @xmath3 .
theorem [ thm : oneend ] implies the following result in hyperbolic geometry : [ cor : identityderived ] suppose that @xmath3 is a complete one - ended orientable hyperbolic surface .
let @xmath58 be a deformation of hyperbolic surfaces such that the derivative of the length satisfies : @xmath59 for all simple closed curves @xmath60 .
then @xmath61 yair minsky pointed out that this corollary can also be deduced from bestvina - bromberg - fujiwara - souto @xcite , using the convexity of length functions with respect to shear deformations .
franois guritaud observed that this corollary also follows from the theory of strip deformations developed in his collaboration @xcite with jeffrey danciger and fanny kassel .
in particular , he described deformations where all the interior simple loops lengthen , and all but one boundary component lengthens .
mcshane @xcite also discusses differentiated length identities and rearrangements of absolutely convergent series .
papadopoulos - theret @xcite explains a construction of thurston s which shortens ( or lengthens ) all closed geodesics ; see also @xcite .
a direct application of the technique of labourie - mcshane @xcite fails , since the cross - ratios of @xcite will not apply directly to the non - reductive group @xmath62 .
however , a generalization of these cross - ratios which includes infinitesimal information may give an alternate proof of our results . as noted above
, mcshane s original identity @xcite concerned once - punctured tori . the generalized mcshane identity for a bordered _ orientable _ hyperbolic surface
@xmath3 is due , independently , to tan - wong - zhang @xcite and mirzakhani @xcite .
norbury @xcite extended the identity to include non - orientable surfaces . in subsequent work
we consider identities for affine deformations when @xmath3 is possibly nonorientable and may have more than one end .
we are grateful to s. antman , a. eskin , f. guritaud , g. mcshane , m. mirzakhani , y. minsky , s.p .
tan and s. wolpert for helpful conversations .
we especially thank m. mirzakhani for the proof of lemma [ lem : mirzakhani ] .
we thank the anonymous referee for several helpful suggestions .
finally we express our appreciation to the mathematical sciences research institute for their hospitality in 2015 when this work was completed .
the fricke space @xmath1 embeds in the quotient space @xmath63 , where @xmath28 acts on the @xmath38-algebraic set @xmath45 by left - composition with inner automorphisms of @xmath28 .
here @xmath64 , and indeed @xmath1 embeds as a connected component of @xmath63 .
openness of @xmath1 in @xmath63 was first proved by weil @xcite .
( see also raghunathan @xcite , theorem 6.19 ) .
it is a special case of the openness of the holonomy map , which was first noticed by thurston ; compare the discussion in theorem 7.2 of kapovich @xcite , goldman @xcite , canary - epstein - green @xcite and koszul @xcite , chapter iv ,
3 , theorem 3 .
closedness is originally due to chuckrow @xcite in this particular context , but it follows from kazhdan - margulis uniform discreteness @xcite ; see chapter viii of @xcite , 4.12 of kapovich @xcite , or 4.1 . of thurston @xcite .
( see also goldman - millson @xcite . )
the fenchel - nielsen coordinates imply that @xmath65 is connected , completing the proof that @xmath1 is a connected component of @xmath63 .
( see , for example , buser @xcite , abikoff @xcite , thurston @xcite,4.6 or wolpert @xcite for details . )
tangent vectors to @xmath1 identifies with affine deformations of fuchsian representations of @xmath66 .
affine deformation _ of @xmath50 consists of a lift of @xmath67 to : @xmath68^{l } \\
\gamma_0 \ar@{-->}[ur]^{\rho } \ar@{^{(}->}[r ] & { \mathsf{so}(2,1)}}\ ] ] where @xmath11 is @xmath10-dimensional _ minkowski space , _ that is , a simply connected geodesically complete flat lorentzian @xmath10-manifold , and @xmath69 is projection onto the linear part .
affine deformations of fuchsian groups arise as holonomy groups of _ margulis spacetimes , _ complete flat lorentz @xmath10-manifolds with free fundamental group .
affine deformations correspond to infinitesimal deformations of the hyperbolic structure on the quotient surface @xmath70 , using the identification of @xmath71 with the _ tangent bundle lie group _ as in .
the infinitesimal deformation theory was developed by weil @xcite ( see also raghunathan @xcite , 6 ) , whereby the tangent space to @xmath1 identifies with the cohomology group of @xmath6 with coefficients in the composition of the holonomy representation with the adjoint representation @xmath72 .
the tangent space to the fricke space at @xmath3 identifies with @xmath73 , where @xmath16 is the image of the holonomy representation for @xmath3 and @xmath74 is the lie algebra of the group of isometries of the hyperbolic plane . since @xmath75
, a tangent vector @xmath76 corresponds to an _ affine deformation _ of @xmath16 , that is , a group of affine isometries of the minkowski spacetime @xmath77 . given @xmath78 , the affine deformation of @xmath2 acts by translations on a unique invariant spacelike line @xmath22 .
the signed lorentzian distance under which this translation displaces is called the _ margulis invariant _
@xmath21 of the affine deformation . a _ deformation _ of hyperbolic surfaces @xmath58 is a one - parameter family of marked hyperbolic surfaces of fixed topology , varying smoothly in @xmath79 $ ] , where @xmath80\in{\mathbb{r}}$ ] is a closed interval .
the space of equivalence classes of marked hyperbolic surfaces forms a smooth manifold , corresponding to a smooth submanifold of the space of representations @xmath81 .
thus a deformation @xmath58 corresponds to a smooth path of holonomy representations @xmath82 in the top - dimensional stratum of @xmath83 .
_ infinitesimal deformations _ of @xmath3 correspond to tangent vectors in this space , and also to affine deformations defined above .
the margulis invariant equals the derivative of the geodesic length function @xmath84 at @xmath85 ( goldman - margulis @xcite ) . by @xcite ,
the @xmath38-valued function @xmath86 on @xmath50 extends to a continuous function on the compact set of _ geodesic currents _ when @xmath50 is convex cocompact , and is therefore bounded .
thus @xmath87 for some @xmath88 and all @xmath89 .
inequality also holds when @xmath50 is only assumed to finitely generated ; for the elementary argument ( which uses trace identities in @xmath90 ) , see @xcite .
let @xmath3 be a complete hyperbolic surface with one end , which we denote @xmath91 .
the length of a closed geodesic @xmath2 in @xmath3 will be written @xmath20 , or simply @xmath23 , when the context is clear .
when @xmath2 is peripheral , then defining @xmath20 to be zero makes the function @xmath92 continuous on the deformation space .
the end of @xmath3 lies in various three - holed spheres such that the other two ends correspond to simple closed geodesics @xmath93 .
( these curves coincide when @xmath94 . )
the set @xmath55 of isotopy classes of such subsurfaces identifies with the set of subsets @xmath95 of isotopy classes of simple closed curves such that @xmath96 cobound a three - holed sphere in @xmath3 .
set : @xmath97 where @xmath98 is the continuous function : @xmath99 the generalized mcshane identity for one - ended orientable @xmath3 is : @xmath100 we briefly discuss the convergence of this series , as the ideas will be crucially used when we differentiate it .
the key idea is that while the number of simple closed geodesics grows polynomially , the summands in these series decay exponentially .
break @xmath55 into a countable disjoint union of finite sets @xmath101 as follows . for every nonnegative integer @xmath102 ,
@xmath103 is finite with cardinality @xmath104 ( mirzakhani @xcite , proposition 3.1 ( 3.15 ) and rivin @xcite(1 ) ) .
compare also rivin @xcite and wolpert @xcite for further discussion .
the constant @xmath105 depends only on the genus @xmath106 and a quantity @xmath107 defined as follows .
choose @xmath108 sufficiently small so that closed geodesics shorter than @xmath39 do not intersect , such as @xmath109 compare buser @xcite , 4 .
the margulis lemma and the collar lemma imply only finitely many such geodesics exist .
explicitly , any set of disjoint closed geodesics extends to a pants decomposition .
let @xmath110 denote the number of ends of @xmath3 .
since the number of pants equals @xmath111 where @xmath112 is the euler characteristic , at most @xmath113 closed geodesics are shorter than @xmath39 .
let @xmath107 be the product of the inverses of the lengths of these geodesics . by taking @xmath114 in an interval @xmath80 $ ] containing @xmath5 ,
the continuity of @xmath107 implies we can choose @xmath115 so that inequality holds for all @xmath79 $ ] although the finite subsets @xmath101 may vary with @xmath114 . to prove that the series converges absolutely , rearrange its terms as follows : @xmath116 where @xmath117 denotes the finite sum of positive terms : @xmath118 @xmath119 let @xmath120 and @xmath121 . write @xmath122 where @xmath123 using the estimate @xmath124 for
@xmath125 , and taking @xmath126 , @xmath127 combine with , and the definition of @xmath128 , obtaining : @xmath129 since @xmath130 , the series for the generalized mcshane identity converges uniformly and absolutely .
now differentiate the generalized mcshane identity to establish uniform absolute convergence of @xmath131 .
the partial derivatives @xmath132 decay exponentially as well : @xmath133 by , the coefficient @xmath137 of @xmath138 and the coefficient @xmath139 of @xmath140 above are each bounded by @xmath141 the contributions from the first term for @xmath135 are bounded by @xmath142 and the contributions from the second terms are bounded by @xmath143 using the linear bound for the margulis invariant . adding these contributions , @xmath144 is bounded by @xmath145 e^-n/2 .
[ prop : identityderivedr ] let @xmath58 be a smooth path of marked complete orientable one - ended hyperbolic surfaces tangent to an affine deformation with margulis invariant @xmath146 .
then : @xmath147 converges absolutely to @xmath138 .
[ proof of corollary [ cor : identityderived ] ] apply proposition [ prop : identityderivedr ] , obtaining : @xmath164 by lemma [ lem : mirzakhani ] , the coefficient on the left - hand side is negative . by
, each coefficient on the right - hand side is negative also . if all @xmath165 for interior curves , then @xmath166 as desired . | we derive an identity for margulis invariants of affine deformations of a complete orientable one - ended hyperbolic surface following the identities of mcshane , mirzakhani and tan - wong - zhang . as a corollary , a deformation of the surface which infinitesimally lengthens all interior simple closed curves must infinitesimally lengthen the boundary .
[ thm]proposition |
the discovery of new models of quantum computation ( qc ) , such as the one - way quantum computer @xcite and the projective measurement - based model @xcite , have opened up new experimental avenues toward the realisation of a quantum computer in laboratories . at the same time
they have challenged the traditional view about the nature of quantum computation . since the introduction of the quantum turing machine by deutsch @xcite , unitary transformations plays a central rle in qc .
however , it is known that the action of unitary gates can be simulated using the process of quantum teleportation based on projective measurements - only @xcite .
characterizing the minimal resources that are sufficient for this model to be universal , is a key issue .
resources of measurement - based quantum computations are composed of two ingredients : ( @xmath0 ) a universal family of observables , which describes the measurements that can be performed ( @xmath1 ) the number of ancillary qubits used to simulate any unitary transformation .
successive improvements of the upper bounds on these minimal resources have been made by leung and others @xcite .
these bounds have been significantly reduced when the state transfer ( which is a restricted form of teleportation ) has been introduced : one two - qubit observable ( @xmath2 ) and three one - qubit observables ( @xmath3 , @xmath4 and @xmath5 ) , associated with only one ancillary qubit , are sufficient for simulating any unitary - based qc @xcite .
are these resources minimal ? in @xcite , a sub - family of observables ( @xmath2 , @xmath4 , and @xmath6 ) is proved to be universal , however two ancillary qubits are used to make this sub - family universal .
these two results lead to an open question : is there a trade - off between observables and ancillary qubits in measurement - based qc ? in this paper , we reply in the negative to this open question by proving that the sub - family @xmath7 is universal using only one ancillary qubit , improving the upper bound on the minimal resources required for measurement - based qc .
the simulation of a given unitary transformation @xmath8 by means of projective measurements can be decomposed into : * ( _ step of simulation _ ) first , @xmath8 is probabilistically simulated up to a pauli operator , leading to @xmath9 , where @xmath10 is either identity or a pauli operator @xmath11 , or @xmath12 . *
( _ correction _ ) then , a corrective strategy consisting in combinig conditionally steps of simulation is used to obtain a non - probabilistic simulation of @xmath8 .
teleportation can be realized by two successive bell measurements ( figure [ fig : telep ] ) , where a bell measurement is a projective measurement in the basis of the bell states @xmath13 , where @xmath14 .
a step of simulation of @xmath8 is obtained by replacing the second measurement by a measurement in the basis @xmath15 ( figure [ fig : gtelep ] ) . up to a pauli operator ]
if a step of simulation is represented as a probabilistic black box ( figure [ fig : blackbox ] , left ) , there exists a corrective strategy ( figure [ fig : blackbox ] , right ) which leads to a full simulation of @xmath8 .
this strategy consists in conditionally composing steps of simulation of @xmath8 , but also of each pauli operator .
a similar step of simulation and strategy are given for the two - qubit unitary transformation @xmath16 ( _ controlled_-@xmath3 ) in @xcite .
notice that this simulation uses four ancillary qubits . as a consequence ,
since any unitary transformation can be decomposed into @xmath16 and one - qubit unitary transformations , any unitary transformation can be simulated by means of projective measurements only .
more precisely , for any circuit @xmath17 of size @xmath18 with basis @xmath16 and all one - qubit unitary transformations and for any @xmath19 , @xmath20 projective measurements are enough to simulate @xmath17 with probability greater than @xmath21 . approximative universality , based on a finite family of projective measurements , can also be considered .
leung @xcite has shown that a family composed of five observables @xmath22 is approximatively universal , using four ancillary qubits .
it means that for any unitary transformation @xmath8 , any @xmath19 and any @xmath23 , there exists a conditional composition of projective measurements from @xmath24 leading to the simulation of a unitary transformation @xmath25 with probability greater than @xmath26 and such that @xmath27 . in order to decrease the number of two - qubit measurements
four in @xmath24 and the number of ancillary , an new scheme called _ state transfer _ has been introduced @xcite .
the state transfer ( figure [ fig : statetrans ] ) replaces the teleportation scheme for realizing a step of simulation .
composed of one two - qubit measurements , two one - qubit measurements , and using only one ancillary qubit , the state transfer can be used to simulate any one - qubit unitary transformation up to a pauli operator ( figure [ fig : statetransgene ] ) .
for instance , simulations of @xmath28 and @xmath29 see section [ sec : unituniv ] for definitions of @xmath28 and @xmath30 are represented in figure [ fig : simulh ] . moreover
a scheme composed of two two - qubit measurements , two one - qubit measurements , and using only one ancillary qubit can be used to simulated @xmath16 up to a pauli operator ( figure [ fig : simulcnot ] ) .
since @xmath31 is a universal family of unitary transformations , the family @xmath32 of observables is approximatively universal , using one ancillary qubit @xcite .
this result improves the result by leung , since only one two - qubit measurement and one ancillary qubit are used , instead of four two - qubit measurements and four ancillary qubits .
moreover , one can prove that at least one two - qubit measurement and one ancillary qubit are required for approximative universality .
thus , it turns out that the upper bound on the minimal resources for measurement - based qc differs form the lower bound , on the number of one - qubit measurements only . and @xmath29 up to a pauli operator.,title="fig : " ] and @xmath29 up to a pauli operator.,title="fig : " ] up to a pauli operator .
] in @xcite , it has been shown that the number of these one - qubit measurements can be decreased , since the family @xmath33 , composed of one two - qubit and only two one - qubit measurements , is also approximatively universal , using _ two _ ancillary qubit . the proof is based on the simulation of @xmath3-measurements by means of @xmath4 and @xmath2 measurements ( figure [ fig : xsimul ] ) .
this result leads to a possible _ trade - off _ between the number of one - qubit measurements and the number of ancillary qubits required for approximative universality .
-measurement simulation ] in this paper , we meanly prove that the family @xmath34 is approximatively universal , using only one ancillary qubit .
thus , the upper bound on the minimal resources required for approximative universality is improved , and moreover we answer the open question of the trade - off between observables and ancillary qubits . notice that we prove that the trade - off conjectured in @xcite does not exist , but another trade - off between observables and ancillary qubits may exist since the bounds on the minimal resources for measurement - based quantum computation are not tight .
there exist several universal families of unitary transformations , for instance @xmath35 is one of them : @xmath36 , @xmath37 , @xmath38 @xmath39 we prove that the family @xmath40 is also approximatively universal : [ thm : dd ] @xmath41 is approximatively universal . the proof is based on the following properties .
let @xmath42 be the rotation of the bloch sphere about the axis @xmath43 through an angle @xmath44 .
[ univer : rotdef ] if @xmath45 is a real unit vector , then for any @xmath44 , @xmath46 .
[ univer : approxrot ] for a given vector @xmath43 of the bloch sphere , if @xmath47 is an irrational multiple of @xmath48 , then for any @xmath44 and any @xmath49 , there exists @xmath50 such that @xmath51 [ univer : decomprot ] if @xmath52 and @xmath53 are non parallel vectors of the bloch sphere , then for any one - qubit unitary transformation @xmath8 , there exists @xmath54 such that : @xmath55 [ univer : irra ] if @xmath44 is not an integer multiple of @xmath56 and @xmath57 , then either @xmath44 or @xmath58 is an irrational multiple of @xmath48 .
_ proof of theorem [ thm : dd ] : _
first we prove that any @xmath59-qubit unitary transformation can be approximated by @xmath29 and @xmath60 .
consider the unitary transformations @xmath61 , @xmath62 , @xmath63 .
notice that @xmath30 is , up to an unimportant global phase , a rotation by @xmath56 radians around @xmath64 axis on the block sphere : [ cols="<,^,<,^ , < " , ] @xmath65 is a rotation of the bloch sphere about an axis along @xmath66 and through the angle @xmath47 .
thus , for any @xmath44 and any @xmath49 , there exists @xmath50 such that @xmath67 since @xmath43 and @xmath53 are non - parallel , any one - qubit unitary transformation @xmath8 , according to proposition [ univer : approxrot ] , can be decomposed into rotations around @xmath43 and @xmath53 : there exist @xmath54 such that @xmath68 finally , for any @xmath8 and @xmath49 , there exist @xmath69 such that @xmath70 thus , any one - qubit unitary transformation can be approximated by means of @xmath71 , and @xmath65 . since @xmath72 and @xmath73 , the family @xmath74 approximates any one - qubit unitary transformation . with the additional @xmath75 gate ,
the family @xmath76 is approximatively universal .
in @xcite , the family of observables @xmath78 is proved to be approximatively universal using two ancillary qubits .
we prove that this family requires only one ancillary qubit to be universal : [ thm : obsmin ] @xmath78 is approximatively universal , using one ancillary qubit only .
the proof consists in simulating the unitary transformations of the universal family @xmath79 .
first , one can notice that @xmath29 can be simulated up to a pauli operator , using measurements of @xmath34 , as it is depicted in figure [ fig : simulh ] .
so , the universality is reduced to the ability to simulate @xmath75 and the pauli operators pauli operators are needed to simulated @xmath80 , but also to perform the corrections required by the corrective strategy ( figure [ fig : blackbox ] ) .
[ lem : simulcz ] for a given 2-qubit register @xmath81 and one ancillary qubit @xmath82 , the sequence of measurements according to @xmath83 , @xmath84 , @xmath85 , and @xmath86 ( see figure [ fig : simulcz ] ) simulates @xmath87 on qubits @xmath81 , up to a pauli operator .
the resulting state is located on qubits @xmath88 and @xmath82 . ] _ proof : _ one can show that , if the state of the register @xmath81 is @xmath89 before the sequence of measurements , the state of the register @xmath81 after the measurements is @xmath90 , where @xmath91 and @xmath92 s are the classical outcomes of the sequence of measurements .
@xmath93 in order to simulate pauli operators , a new scheme , different from the state transfer , is introduced .
[ lem : simulz ] for a given qubit @xmath94 and one ancillary qubit @xmath88 , the sequence of measurements @xmath95 , @xmath96 , and @xmath95 ( figure [ fig : simulz ] ) simulates , on qubit @xmath94 , the application of @xmath97 with probability @xmath98 and @xmath99 with probability @xmath98 . , scaledwidth=50.0% ] _ proof : _ let @xmath100 be the state of qubit @xmath94 .
after the first measurement , the state of the register @xmath81 is @xmath101 where @xmath102 is the classical outcome of the measurement .
since @xmath103 , the state of the register @xmath81 is : @xmath104 let @xmath105 be the outcome of the last measurement , on qubit @xmath88 .
if @xmath106 then state of the qubit @xmath94 is @xmath89 , and @xmath107 otherwise .
one can prove that these two possibilities occur with equal probabilities .
[ lem : simulx ] for a given qubit @xmath94 and one ancillary qubit @xmath88 , the sequence of measurements @xmath108 , @xmath109 , and @xmath108 ( figure [ fig : simulx ] ) simulates , on qubit @xmath94 , the application of @xmath110 with probability @xmath98 and @xmath99 with probability @xmath98 .
, scaledwidth=50.0% ] the proof of lemma [ lem : simulx ] is similar to the proof of lemma [ lem : simulz ] .
_ proof of theorem [ thm : obsmin ] : _ first notice that the family of unitary transformations @xmath111 is approximatively universal since @xmath112 is universal .
@xmath29 and @xmath113 can be simulated up to a pauli operator ( lemmas [ lem : simulcz ] ) .
the universality of the family of observables @xmath78 is reduced to the ability to simulate any pauli operators .
lemma [ lem : simulx ] ( resp .
lemma [ lem : simulz ] ) , shows that @xmath110 ( @xmath97 ) can be simulated with probability @xmath98 , moreover if the simulation fails , the resulting state is same as the original one .
thus , this simulation can be repeated until a full simulation of @xmath110 ( @xmath97 ) . finally , @xmath114 can be simulated , up to a global phase , by combining simulations of @xmath115 and @xmath97 .
thus , @xmath78 is approximatively universal using only one ancillary qubit . @xmath93
we have proved a new upper bound on the minimal resources required for measurement - based qc : one two - qubit , and two one - qubit observables are universal , using one ancillary qubit only .
this new upper bound has experimental applications , but allows also to prove that the trade - off between observables and ancillary qubits , conjectured in @xcite , does not exist .
this new upper bound is not tight since the lower bound on the minimal resources for this model is one two - qubit observable and one ancillary qubit . | we improve the upper bound on the minimal resources required for measurement - based quantum computation @xcite .
minimizing the resources required for this model is a key issue for experimental realization of a quantum computer based on projective measurements .
this new upper bound allows also to reply in the negative to the open question presented in @xcite about the existence of a trade - off between observable and ancillary qubits in measurement - based quantum computation . |
we have recently developed a theory for particle - hole states of closed - shell atoms that is potentially able to predict energies and transition rates with high precision .
initially , this theory has been tested in neon @xcite . for transitions between neon excited states
the accuracy of experiments is very good and experiments agree well with our theory .
however , for heavier noble gas atoms , experiments are often prone to systematic errors and the precision is much lower .
for example , many absolute transition rates of argon were based on high - pressure arc plasmas and depended on the assumption of local thermodynamic equilibrium of the upper state populations which often was satisfied very approximately ( see discussion in ref .
lifetime measurements using different techniques or even the same technique often yield quite different results . in ref .
@xcite , a unified set of atomic probabilities for neutral argon is given .
the data are obtained from extensive analysis of existing experiments and from new measurements of emission in the broad range 240 - 18000 nm .
the data are included in the nist data base .
errors of order 10% are quoted .
although our theory is able to reproduce nist data , the difference between experiment and theory often exceeds quoted in ref .
@xcite error bars .
there are two possibilities : ( i ) the theory is less accurate and needs to be improved , ( ii ) experimental error bars are larger and critical , more precise measurements are needed .
therefore , the accurate measurements presented in this paper are well motivated .
we describe a new experimental technique for precise measurements of ratios of transition rates in gas discharges and detail an apparatus realizing this technique .
we demonstrate our measurement of the ratio of two argon transitions and analyze systematic errors
. finally , we present a comparison with our theory and other theories and experiments .
the essential goal of the experimental technique is to measure the relative absorption rates for two different transitions that start from the same level . in general , the populations of the lower and upper states are needed for finding the cross - section @xmath0 from the absorption coefficient @xmath1 , @xmath2 \sigma _ { if}\ ] ] ( here @xmath3 and @xmath4 are densities of atoms in the initial and final states , @xmath5 is the ratio statistical weights ) , but in some cases the upper state population can be neglected .
this happens when the initial state is the ground or long - lived metastable state , and the final , upper states are short - lived and relatively unpopulated . in gas discharges , at low current densities , metastable levels are generally much more populated than any other non - metastable excited levels .
experimentally , we also observed states that have allowed decay modes .
the @xmath6 _ { 1}$ ] and @xmath7 _ { 1}$ ] states , surprisingly , were populated almost as much as metastables at pressure of a few torr .
the observed lifetimes of these states are of the order of ten microseconds .
the principal explanation for such a slow decay is radiation capture : many possibilities of emission and absorption of a photon occur before the photon can leave the discharge cell .
the observation of long - lived species in discharges gives the possibility to use them as lower states for absorption measurements .
the other excited levels ( not belonging to the 4s group ) quickly decay radiatively without radiation capture and are effectively depopulated in a pulsed discharge as soon as the discharge current disappears .
therefore , in upward transitions that start from the 4s levels , the condition @xmath8 is satisfied , and the absorption becomes proportional just to the cross - section and the initial density .
thus , by measuring the absorption for two transitions starting from the same level , in our experiment @xmath7 _ { 1}$ ] , we can find the ratio of cross - sections from the ratio of absorptions , @xmath9 for a doppler broadened profile and for monochromatic laser light tuned to the maximum of the profile , the ratio of cross - sections is equal to the ratio of line strengths @xmath10 therefore by measuring the ratio of absorptions for two transitions , we can find the ratio of line strengths @xmath11 , the important parameter that can be compared with other experiments and with theory .
the situation described above is ideal for precise measurements of transition rate ratios : it has been very closely reproduced in alkali vapor experiments , in which for cs an accuracy of better than 0.1% is claimed by @xcite . however , in the case of noble gases , especially when we choose the starting level to be a non - metastable state , a more complicated analysis is required . in order to satisfy the condition @xmath8 for the lowest excited states , we use a pulsed rather than a dc discharge . in a dc discharge
this condition is weaker , because radiatively decaying levels are constantly repopulated by collisions with hot electrons and by decay of higher levels .
it is especially true for radiation - capturing levels .
a further advantage of a pulsed discharge is the possibility for enhancing the signal by utilizing its time dependence .
for example slow fluctuations in laser intensity can be excluded .
in addition , the signal - to - noise ratio can be improved using averaging ( 128-scan in our experiment ) on a digital oscilloscope . without special efforts
an absorption of order 0.1% can be easily detected .
such high sensitivity is useful for fast and precise tuning of the laser to centers of doppler profiles which is needed if only one laser is available .
the long - term stability of the pulsed discharge is also better : the pulsed discharge has substantially smaller sputtering and thermal effects , since the total charge from the discharge current is much smaller for a pulsed discharge .
the principal disadvantage of the pulsed discharge is the more complicated analysis of systematic errors , basically owing to the finite - time response of the detection system .
the various factors that can influence the measurement of the ratio in pulsed discharges need careful analysis : we will consider systematic errors that arise from large absorption and inhomogeneity .
when two transitions differ by a factor of at least three , a larger absorption is recommended for better signal - to - noise ratio .
in such a case the linear relation between absorption and line strength may no longer hold , and instead beer s law must be used to derive the more general equation : @xmath12 here @xmath13 and @xmath14 are the reductions in the intensity after passing the absorber for the first and the second transitions , respectively .
this nonlinear equation can be the cause of several systematic errors .
one effect arises from a slow response of the detection system .
this results in a reduction of actual height of the peak by an unknown coefficient @xmath15 : @xmath16 if the equations were linear , this coefficient would drop out from the ratio , but for large and significantly different absorptions for the two transitions , it will cause a systematic error in the ratio .
the other effect arises from inhomogeneity .
so far we assumed that the absorption medium is homogeneous , but in gas discharges this is far from being true .
the discharge produces densities of excited atoms with some transverse and longitudinal distributions .
the longitudinal distribution does not cause any problem , since for two transitions the ratio at each small length of path stays the same , and eventually the ratio of logarithms of intensities in eq .
( [ ratiolog ] ) will remain the same . the transverse distribution , however , will cause a deviation from the simple beer s law and will require the analysis of this distribution . for a cylindrical discharge cell , which we use in our experiment ,
if @xmath17 is the distance from the axis , an integration over @xmath17 is needed , @xmath18 2\pi rdr\text{,}\ ] ] where @xmath19 is the input laser beam intensity , @xmath20 is the absorption coefficient , and @xmath21 is integrated over a detector area output intensity after an inhomogeneous discharge medium . for simplicity in this equation
the change in absorption along the laser beam is ignored .
it is difficult to find the spatial distribution accurately , although in some simple cases it can be predicted . in the case of a diffusion controlled cylindrical discharge ,
the distribution for the electron density @xmath22 and for excited state densities , has approximately the form of a bessel function , @xmath23 the expression for the electron density @xmath22 is taken from ref .
@xcite , p. 67 . using the series expansion of the bessel function for small argument ,
@xmath24 we can show that in our experiment with @xmath25 cm and @xmath26 cm , the maximum change in the absorption over the beam radius @xmath17 is 3.7% and the change in the integral for small @xmath20 is only 1.8% .
therefore the effect of the distribution is small .
moreover , if the absorption is not very large , the exponential can be expanded , and the linear dependence of the signal on the absorption coefficient can reduce further the effect of the distribution . if we examine nist transition data @xcite for neutral argon , we find many transitions in the range of tunable lasers such as a ti : sapphire laser , which can be used for the absorption experiments .
however , in practice there are some restrictions : first , the pair of transitions has to start from the same metastable levels or radiation - capturing levels such as mentioned above .
second , the range is greatly restricted by the tunability of the laser : for example , our ti : sapphire laser has several mirror sets for several ranges of wavelengths . because the change of mirrors takes many hours to complete , only one mirror set can be used for measurements of ratios at one time . as a result , for our currently installed mirror set , only two pairs of transitions were appropriate . however
, our laser could not be tuned for one transition , and we ended up able to make measurements for only one pair of transitions : one transition at a wavelength 9227 and the other at 9787 .
the optical and electrical diagrams of the experimental arrangement are shown in fig .
[ figa1 ] . the monochromatic 1w laser beam generated by the ti :
sapphire laser is expanded by traveling 8 meters between two optical tables to the size of about 5 cm diameter .
a variable aperture reduces both the beam s size and its power . to avoid saturation of the transition
, the beam intensity is further reduced using a 2% reflection off a glass plate .
a chopper periodically interrupts the beam to give a 100% reference absorption .
the beam is then directed through the discharge cell where it undergoes time - dependent absorption in the discharge initiated by hv pulses from the hv pulse generator .
the optical path of the laser beam is terminated on the photodiode .
the attenuator in front of the screened photodiode compartment is used to avoid any saturation effects in the detection system .
the signal from the photodetector is amplified and sent to the digital oscilloscope for measurement of the absorption .
the laser is a commercial single frequency mbr-110 ti : sapphire laser , pumped with 7 - 15 w of cw coherent argon ion laser .
the mbr-110 laser can be tuned in a broad range , 700 - 1060 nm , but to cover all this range 6 different mirror combinations have to be installed .
the discharge cell is a quartz tube of diameter 2.5 cm and a length of 30 cm .
two standard electrodes were fused symmetrically in side - tubes near each end .
the electrodes have large area to reduce unwanted sputtering from the cathode which can reduce the lifetime of metastables significantly and affect the population distribution in the discharge as we observed in detail in helium .
moreover , for the same purpose , an external ballast volume was added to the cell and a short - pulsed discharge was used instead of a dc discharge .
a high voltage pulse generator produces high - voltage ( more than 10 kv to allow fast discharge breakdown ) pulses each of a few microseconds .
the circuit is based on a power transistor in the switching mode and high - voltage transformer .
a rotating chopper is inserted into the optical path to produce a 100% normalization signal .
the detection system consists of a photodiode , an amplifier , and a digital oscilloscope .
there is no special requirement on those elements except that the detection system should provide a linear response and the distortion of the waveform of the signal should be minimal . the first requirement is completely fulfilled if the laser intensity does not exceed the saturation limit for a particular diode .
the second requirement is approximately satisfied : small distortion of the signal has been observed .
this effect is analyzed in detail in the next section .
to find the ratio of line strengths , we measure the peak absorption for two transitions and use the formula @xmath27 here @xmath28 and @xmath29 are the voltages proportional to the maximum of a pulsed discharge absorption and to the chopper normalization absorption for the transition at @xmath30 nm shown in fig.[figa3 ] .
similar voltages for the second transition at @xmath31 nm also shown in fig.[figa3 ] .
are @xmath32 and @xmath33 . repeating measurements of absorption several times
, we have obtained the ratio @xmath34 .
a small error here indicates the excellent statistics .
however , the result can still have systematic errors which need to be estimated and if possible be excluded .
the first type of systematic error can occur from a slow drift in discharge conditions . though our pulsed discharge is very stable , to account for any slow drift , we alternate measurements of the absorptions for the two transitions .
the second possible systematic error can arise from the uncertainty in the wavelength when the laser is manually tuned to the resonances of the transitions by monitoring absorption peaks similar to shown in fig .
[ figa3 ] . if @xmath35 and @xmath36 are relative average shifts owing to tuning uncertainty , then the absolute systematic shift in the ratio @xmath37 is @xmath38 } { \ln \left [ \left ( 1-\left ( 1-\varepsilon _ { 2}\right ) v_{2}/v_{20}\right ) \right ] } -\frac{s_{1}% } { s_{2}}\approx ( \varepsilon _ { 2}-\varepsilon _ { 1})\text{,}\ ] ] where this equation was simplified for small absorption .
the tuning error is estimated to be about 1.5% .
the combined effect of the nonlinearity of absorption and finite - time response of the detection system leads to another systematic error .
suppose @xmath39 is the reduction in the peak height because of a slow time response .
then we can write for the @xmath40-ratio @xmath41 } { \ln % \left [ 1-\left ( 1-\varepsilon \right ) v_{2}/v_{20}\right ] } = \frac{s_{1}}{% s_{2}}+\sigma\ ] ] comparing measurements of ratios at different discharge currents and , hence absorption , we find that the time responce error is about @xmath42 .
this error gives a systematic increase in the ratio by 4% , but the previous error due to tuning decreases the ratio slightly by 1.5% so that the overall systematic error is smaller . finally , taking into account mentioned above statistical and systematic errors
, we obtain the accuracy of the measurement of the ratio about 4% .
the final value of our ratio is @xmath43 .
.the ratio of the two line strengths at 922.7 nm and 978.7 nm [ cols="<,<",options="header " , ] the comparison of the ratio of line strengths is given in table [ argtb ] .
our measured ratio of line strengths @xmath44 is in excellent agreement with the value 3.23 obtained from our mixed configuration - interaction and perturbation - theory calculations @xcite . both measured and theoretical ratios agree also with the ratio 2.89@xmath4515% calculated from the experimental transition rates of ref .
our results are close to semiempirical calculations by @xcite . however , the ratio 4.62 from semiempirical calculations of ref .
@xcite based on a similar principle as calculations in ref .
@xcite disagree substantially with our result .
one obvious problem with semiempirical approaches is that energies can not define completely the wave functions which are very sensitive to the accuracy of calculations .
in addition , transition amplitudes have additional corrections which are not accounted by the effective hamiltonian . for example , random - phase approximation ( rpa ) corrections are due to shielding by the atom of an external photon field .
those corrections are substantial in ar and heavier noble - gas atoms and must be included to reach a good precision .
we have measured an accurate ratio of line strengths for two argon transitions . for this measurement we developed a new experimental technique based on a pulsed discharge and a laser absorption measurement .
this technique can be applied for measurements of other ratios in noble gas atoms .
the agreement of our measurement and calculations is excellent . | the ratio of two line strengths at 922.7 nm and 978.7 nm of argon is measured in an argon pulsed discharge with the use of a single - mode ti : sapphire laser .
the result 3.29(0.13 ) is in agreement with our theoretical prediction 3.23 and with a less accurate ratio 2.89(0.43 ) from the nist database . |
the 17 m diameter magic telescope on the canary island of la palma is the world s largest single imaging atmospheric cherenkov telescope .
one of the aims of the magic collaboration is to carry out observations at large zenith angles . under these special conditions and with a good telescope sensitivity a bigger effective area
is given and sources from a large section of the southern sky can be observed with a threshold of a few hundred gev ( @xmath2100gev-500gev , zenith angle dependent ) . here
we present the results of an analysis of the crab nebula data set taken under large zenith angles ( 60@xmath3 to 66@xmath3 ) .
in addition to the usual image parameters @xcite timing information of the recorded signals was used for reconstruction of the shower origin and for the background suppression .
furthermore we present a reanalysis of a data set recorded at a zenith angle range between 59@xmath3 and 64@xmath3 . the same analysis and cuts are used as for the crab nebula data set .
all the data analysed in this work was taken in wobble mode , i.e. tracking a sky direction , which is 0.4@xmath3 off the source position .
the background is estimated from the same field - of - view , which improves the background estimation and yields a better time coverage because no extra off data have to be taken .
compared to the previous analysis @xcite improvements are obtained because of an updated monte carlo ( mc ) sample at high zenith angles leading to a better agreement . a further improvement is achieved in the image cleaning and in the gamma / hadron separation thanks to the usage of the timing information of the images . in the new analysis
we use the time image cleaning @xcite : with a sub - nsec timing resolution of the data acquisition system and thanks to the parabolic structure of the telescope mirror a smaller integration window can be chosen .
this reduces the number of pixels with signals due to night sky background , which survive the image cleaning .
this allows reducing the pixel threshold level ( i.e. recorded charge ) of the image cleaning leading to a lower analysis energy threshold . for the analysis a robust set of dynamical cuts with a small number of free parameters
is used @xcite .
in addition , cuts in time parameters are applied , which describe the time evolution along the major image axis and the rms of the time spread .
these two additional parameters lead to a better background suppression yielding a better sensitivity of the analysis .
the energy estimation is done with the random forest regression method @xcite .
the crab nebula is one of the best studied celestial objects because of the strong persistent emission of the nebula over 21 decades of frequencies .
it was the first object that was detected at tev energies by the whipple collaboration @xcite in the year 1989 and is the strongest steady source of vhe @xmath1-rays . due to the stability and the strength of the @xmath1-ray emission the crab nebula
is generally considered the standard candle of the tev astronomy .
the measured @xmath1-ray spectrum extends from 60gev @xcite up to 80tev @xcite and appears to be constant over the years ( from 1990 to present ) . in october 2007 ,
the magic telescope took data of the crab nebula with a zenith angle range of 60@xmath3 up to 66@xmath3 .
the data was taken under dark sky conditions and in wobble mode .
after quality cuts an effective on - time of 2.15hrs is obtained . using detection cuts ( i.e. optimized on significance of a different crab nebula sample ) ,
a total of 247 excess events above 187 background events with a scale factor of 0.33 have been detected ( see fig . [
fig : crab_theta ] ) . according to li&ma formula 17 @xcite
the significance of this signal is 12.8@xmath4 .
this corresponds to an analysis sensitivity of 8.7 @xmath5 . using the same set of cuts we obtain the following sensitivities for integral fluxes ( @xmath6 ) : @xmath7 the on - source and normalized background distribution of @xmath8 .
the on - source is shown in the black crosses and the background is shown in the gray shaded region .
2.22hrs of crab data show an excess with a significance of 12.8@xmath4.,scaledwidth=49.0% ] the differential energy spectrum can be described well by a power law : @xmath9 the spectrum is shown in fig . [
fig : crab_spec ] .
the gray band represents the range of results obtained by varying the total cut efficiency between 40% and 70% . for comparison , the crab nebula spectrum from data taken at low zenith angles
is drawn as a dashed line @xcite .
a very good agreement has been found .
differential energy spectrum of the crab nebula .
black line : power law fit to the data , gray band : systematic uncertanties , dashed line : published data taken at low zenith angles @xcite.,scaledwidth=49.0% ] like the crab nebula being the so - called standard candle of @xmath1-ray astronomy , the blazar pks2155 - 304 is the so - called lighthouse of the southern hemisphere .
the high frequency peaked bl lac pks2155 - 304 at a redshift of z=0.116 was discovered in the vhe range by the university of durham mark 6 @xmath1-ray telescope ( australia ) in 1997 with a flux corresponding to @xmath00.2 times the crab nebula flux @xcite . it was later observed and detected with high significance by the southern observatories cangaroo @xcite and h.e.s.s .
@xcite establishing this source as be the best studied southern tev blazar .
detection from the northern hemisphere is difficult due to challenging observation conditions under large zenith angles . in july 2006 , the h.e.s.s .
collaboration reported an extraordinary outburst of vhe @xmath1-emission @xcite . during this outburst ,
the @xmath1-ray emission was found to be variable on time scales of minutes with a mean flux of @xmath07 times the flux observed from the crab nebula .
follow up observations of the outburst by the magic telescope have been triggered in a target of opportunity program by an alert from the h.e.s.s .
collaboration @xcite .
the magic telescope observed the blazar from 28 july to 2 august 2006 in a zenith angle range from 59@xmath3 to 64@xmath3 .
the data were taken under dark sky conditions and in wobble mode .
after quality cuts a total effective on - time of 8.7hrs is obtained . for the detection of ,
the same cuts are used as for the detection of the crab nebula .
three off regions are used and 1029 excess events above 846 background events are detected .
a significance of 25.3 standard deviations is obtained .
the corresponding @xmath8-plot is presented in fig .
[ fig:2155_theta ] .
the on - source and normalized background distribution of @xmath8 .
the denotations are the same as in figure [ fig : crab_theta ] .
a clear excess with a significance of more than 25 standard deviations for a source at the position of pks2155 - 304 is found.,scaledwidth=49.0% ] the differential energy spectrum is shown in fig .
4 as a black line together with the measured spectrum of h.e.s.s . during the strong outburst @xcite ( dashed line ) . note that h.e.s.s . and
magic data are not simultaneous .
the obtained spectral points in this analysis are fitted from 400gev on , because at lower energies h.e.s.s . reported a change of the slope ( @xmath10 above 400gev to @xmath11 below 400gev ) .
the fitted magic data points are consistent with a power law : @xmath12 with a fit probability after the @xmath13-test of 81% . above 400gev ,
the energy spectrum measured by h.e.s.s . from the preceding flare of pks2155 - 304 is one order of magnitude higher than the spectrum measured by magic , but the spectral slope ( @xmath10 ) is consistent within the statistical errors .
measured and intrinsic differential energy density of pks2155 - 304 .
the effect of the ebl is taken into account by using the recent model of kneiske adopted by magic @xcite .
the power law fit to the observed spectrum is given by the dotted line , while the curved fit to the intrinsic spectrum is shown by the solid line .
the fitted peak position is located at @xmath14 . for comparison
the crab nebula spectrum is shown as the dashed line.,scaledwidth=49.0% ] the vhe photons of pks2155 - 304 interact with the low - energy photons of the extragalactic background light ( @xcite ) .
the predominant reaction @xmath15 leads to an attenuation of the intrinsic spectrum @xmath16 that can be described by @xmath17 \nonumber\end{aligned}\ ] ] with the observed spectrum @xmath18 , and the energy dependent optical depth @xmath19 . here
we use the recent model of kneiske et al .
that has been adopted by magic @xcite . the measured spectrum and the reconstructed deabsorbed spectrum
are shown in fig .
[ fig:2155_deab ] . for comparison
the crab nebula spectrum is also shown .
a power law fit to the deabsorbed spectrum results in a spectral index of @xmath20 .
however , the fit probability is rather low ( 4% ) which motivates a higher order fit function .
we have chosen a curved power - law fit ( a parabolic shape in log - log representation ) of the following form : @xmath21 .
the best fit parameters are : , @xmath22 ( solid line in fig .
[ fig:2155_deab ] ) and the fit probability is 77% .
the observed curvature is indicative of a maximum in the energy density and is usually interpreted as due to inverse compton ( ic ) scattering .
the fitted peak position is determined to be at @xmath14 .
the integral light curves above 400gev shown in fig .
[ fig : lc ] have a binning of one flux point per night ( bottom panel ) and a binning in two runs which corresponds to about 10minutes per bin ( top panel ) .
significant detections in most of the time bins are obtained .
a significant intranight variability is found for the second night mjd53945 ( 29 july 2006 ) giving a probability for a constant flux of less than @xmath23 .
for the other nights , no significant intranight variability is found . in the lower panel of fig .
[ fig : lc ] , a night - by - night light curve is shown .
a fit by a constant to the run - by - run light curve results in a chance probability of less than @xmath24 .
however , a fit by a constant to the night - by - night light curve results in a chance probability of @xmath25 . we , therefore , conclude that there is a significant variability on the time scales reaching from days ( largest scale we probed ) down to 20 minutes ( shortest scale we probed ) .
a study of high zenith angles ( 60@xmath3 - 66@xmath3 ) observations with the magic telescope was performed .
a new time image cleaning and also time parameters were used for the background suppression , which leads to a significant improvement of the sensitivity . from crab nebula observations a sensitivity of 5.7% of the crab nebula flux for 50hrs of observations above 0.4tev has been determined .
the differential energy spectrum of the crab nebula is in excellent agreement with the published data at lower zenith angles .
this improved analysis is used to reanalyze data of taken with magic in 2006 . the energy spectrum from 400gev up to 4tev
has a spectral index of ( @xmath26 ) and shows no change of spectral slope with flux state .
it agrees well with the results of h.e.s.s . @xcite and
cangaroo @xcite .
furthermore we corrected the measured spectrum for the effect of the ebl absorption using the recent model of kneiske et al .
the resulting intrinsic spectrum shows a clear curvature .
a fitted peak position in the energy density distribution is at @xmath14 .
the light curves show a significant variability on daily as well as on intra - night time scales .
finally we conclude that high zenith angle observations with the magic telescope have proven to yield high quality spectra and light curve at a low energy threshold .
we would like to thank the instituto de astrofisica de canarias for the excellent working conditions at the observatorio del roque de los muchachos in la palma . the support of the german bmbf and mpg , the italian infn and spanish micinn is gratefully acknowledged .
this work was also supported by eth research grant th 34/043 , by the polish mniszw grant n n203 390834 , and by the yip of the helmholtz gemeinschaft .
d.ms research is supported by a marie curie intra european fellowship within the 7th european community framework programme .
p. m. chadwick et al .
, _ apj _ 513:161 - 167 , 1999 .
f. a. aharonian et al .
( h.e.s.s .
coll . ) , _ apj _ 664:l71-l74 , 2007 .
a. m. hillas , in proc .
ray conf . , la jolla , usa , 3:445 - 448 , 1985 .
d. mazin and e. lindfors for the magic coll .
, in proc 29th int .
ray conf . ,
merida , mexico , 3:1033 - 1036 , 2008 .
e. aliu et al . , _ astropart .
_ , 30:293 , 2009 . b. riegel and t. bretz , in proc 29th int .
ray conf . , pune , india , 4:315 , 2005 .
j. albert et al .
( magic coll . ) , _ nima _ 588:424 - 432 , 2008 .
t. weekes et al .
, _ apj _ , 342:379 - 395 , 1998 .
j. albert et al .
( magic coll . ) , _ apj _ , 674:1037 - 1055 , 2008 . f. a. aharonian et al .
( hegra coll . ) , _ apj _ , 539:317 - 324 , 2000 .
t. p. li and y. q. ma , _ apj _ , 272:317 - 324 , 1983 .
sakamoto et al .
( cangaroo coll . ) , _ apj _ , 676:113 - 120 , 2008 . f.
a. aharonian et al .
( h.e.s.s .
coll . ) , _
, 430:865 - 875 , 2005 .
r. j. gould & schreder , _ phys.rev .
_ , 16:252 , 1966 .
m. g. hauser & e. dwek , _ ara&a _ , 39:249 , 2001 . t. kneiske , k. mannheim and d. hartmann , _
a&a _ , 386:1 , 2002 .
e. aliu et al .
( magic coll . ) , _ science _ , 320:1752 , 2008 . | the high frequency peaked bl lac with a redshift z=0.116 was discovered 1997 in the vhe range by the university of durham mark 6 telescope in australia with a flux corresponding to @xmath00.2 times the crab nebula flux @xcite . it was later observed and detected with high significance by the southern observatories and h.e.s.s . establishing this source as the best studied southern tev blazar .
detection from the northern hemisphere was very difficult due to challenging observation conditions under large zenith angles .
in july 2006 , the h.e.s.s .
collaboration reported an extraordinary outburst of vhe @xmath1-emission @xcite . during the outburst ,
the vhe @xmath1-ray emission was found to be variable on the time scales of minutes and at a mean flux of @xmath07 times the flux observed from the crab nebula @xcite .
the magic collaboration operates a 17 m imaging air cherenkov telescope at la palma ( northern hemisphere ) .
follow up observations of the extraordinary outburst have been triggered in a target of opportunity program by an alert from the h.e.s.s .
collaboration . the measured spectrum and light curve are presented .
bl lacertae objects : individual ( ) gamma rays : observations methods : data analysis |
the reduction of size in the electronic devices to nanometer scale has highlighted the importance of the effects of quantum coherence and interference .
the control of such effects is important to provide both a better understanding of the quantum realm as well as new functionalities to the circuits @xcite .
quantum interference allows to enhance or to cancel , total or partially , the response of the system beyond the simple classical additive behaviour .
such an effect can pose a problem to be avoided but also could provide new capabilities to the device with respect to its classical counterpart @xcite .
there is currently an increasing interest in being able to tune the parameters of mesoscopic systems , what would enable to manipulate their quantum behaviour for the advantageous design and applications of future electronic devices @xcite .
the characteristic quantum phenomenon of the change of phase of the wave function along two paths enclosing a magnetic flux , i.e. the aharonov - bohm ( ab ) effect @xcite , has been envisioned for the feasible exploitation of the quantum phase in electronic devices @xcite .
phase coherent effects in ab rings have been treated theoretically in the literature @xcite .
closely related to the concept of coherence is the quantum interference between a discrete state with a continuum of states .
this phenomenon was firstly studied by fano in the spectrum of photoionization of atoms and termed fano effect after him @xcite , but was found ubiquitous in a wide range of physical phenomena .
interestingly , it has been observed in properly tailored nanoscale systems @xcite .
the first tunable experiment showing the characteristic asymmetric fano profile in the electron transmission was reported in an ab ring with a quantum dot embedded in one of its arms @xcite .
various parameters , such as the gate voltage and the magnetic flux through the ring among others , allowed to tune the peak and dip of the profile .
more recently , a quantum dot molecule , i.e. two coupled quantum dots , has been embedded between the arms of an ab interferometer , with even a larger number of tunable parameters @xcite ; the transmission thus exhibits a large variety of behaviours as a function of them .
in @xcite it is stated that the simplest fully coherent single - mode picture is of limited use for the interpretation of their experiments .
in such a picture , the transmission is assumed to arise from a direct reflection and another one after travelling once around the ab ring .
it is the purpose of this paper to improve the theoretical description of the experimental results .
+ here we present theoretical calculations of a model inspired in the system of ref .
@xcite within a phase coherent non - interacting formalism .
on one hand , we assessed the sensitivity of the conductance upon variation of the various model parameters . on the other hand , we show that the self - energy matrix elements , obtained from applying a partitioning technique to the hamiltonian , can give insight on the quantum interference and its dependence on the magnetic flux through the ring .
this interpretation in terms of contributions through pathways , allows one to predict the onset or cancellation of the antiresonances as a function of the parameters of the model . +
this paper is organized as follows : section [ model ] describes the model of the device and its connections to the leads .
section [ section partitioning ] introduce the spacial contributions to the conductance by means of a partitioning technique and discusses the conditions for the onset of peaks and cancellation of conductance in the transmission in terms of the green functions of the isolated device . in section [ section results ]
we present the results obtained varying the various model parameters ; finally , in section [ section conclusions ] we summarize our conclusions .
we consider four quantum dots forming a ring and coupled to two leads l and r. sites 2 and 4 of the ring are connected to each other forming an artificial molecule , as shown in fig .
[ system ] .
scheme of the device : ( a ) in the ( 1,2)-connection and ( b ) ( 1,3)-connection to the leads . ]
we consider only one energy level in each dot and both the intradot and interdot electron - electron interactions are neglected .
the system shall be described by a hamiltonian @xmath0 where @xmath1 is the hamiltonian of the isolated bicyclic ring , @xmath2 @xmath3 is the hamiltonian of the leads @xmath4 and @xmath5 is the hamiltonian describing the tunneling between the leads and the ring @xmath6 where @xmath7 are the on - site energy at the dots , @xmath8 are the nearest - neighbour hopping parameters ( where @xmath9 should be understood ) , @xmath10 is the interdot hopping that couples the upper and lower arms of the interferometer , @xmath11 is the phase @xmath12 acquired due to interdot hopping in a magnetic field threading the ring with a reduced flux @xmath13 ( i.e. , in units of the quantum @xmath14 ) , and @xmath15 is the site of contact to the right lead .
the left lead is always attached to the dot 1 , as shown in fig .
[ system ] ; for brevity we refer to them as connections ( 1,2 ) and ( 1,3 ) .
the current through the device can be calculated with the landauer equation @xmath16,\ ] ] where @xmath17 and @xmath18 are the fermi distributions at the l and r leads . at low temperatures ,
the transmission function represents the dimensionless conductance ( in units of the quantum @xmath19 ) and is calculated as @xmath20 where @xmath21 and @xmath22 are the matrix representation of the advanced and retarded green functions , and @xmath23 and @xmath24 are the spectral densities of the leads .
+ in the _ wide band _ approximation , the green function of the connected system is given by @xmath25 where @xmath26 is the retarded green function of the isolated system , and @xmath27 are the self - energies of the leads , considered to be energy independent .
partitioning of the basis space is usually employed for isolating the effects on the part of interest from the rest of the system @xcite . here , we apply a partitioning technique to recover the notion of spatial transmission pathways @xcite .
the @xmath28 hamiltonian can be partitioned in terms of @xmath29 matrices as follows : @xmath30 where @xmath31 is the part of the hamiltonian projected on the subspace of orbitals centered on the sites of connection 1 and @xmath15 , where @xmath32 , whilst @xmath33 is the projection of @xmath34 on the complementary subspace @xmath35 .
matrices @xmath36 and @xmath37 contain matrix elements connecting states belonging to @xmath38 and @xmath39 .
the green function can be obtained by block matrix decomposition @xmath40 we are interested here in the green function projected on the subspace of the connection sites , i.e. , its @xmath38-block .
hence , @xmath41 can be obtained from the inverse of the schur complement of the @xmath39-block , @xmath42 , @xmath43 from which an effective hamiltonian can be defined as @xmath44 where @xmath45 is the self - energy that contains the interactions involving the orbitals not connected to the leads .
an approach similar to the one outlined above has been used in an analytical treatment of quantum interference in a benzene ring @xcite . the green function written in terms of the self - energy ,
@xmath46 with @xmath47 , contains all the matrix elements needed for the calculation of the conductance between the sites 1 and @xmath15 .
their poles are given by the zeroes of the secular determinant @xmath48 , while the antiresonances comes from zeroes in @xmath49 .
when @xmath50 , as in the ( 1,3 ) connection , the antiresonances arise from the zeroes of @xmath51 .
interestingly , the self - energy @xmath52 becomes a sum of contributions throughout paths from above ( a ) , from below ( b ) and through the interarm coupling ( c ) .
the vanishing of the transmission occurs when the contributions from those paths interfere destructively thus cancelling the element @xmath53 of the effective hamiltonian .
+ for the ( 1,3 ) connection , the contributions become @xmath54 where @xmath55 and @xmath56 , with @xmath57 and @xmath58 being the energies of the bonding ( @xmath58 ) and antibonding ( @xmath57 ) orbitals of the molecule formed between the sites 2 and 4 described by @xmath33 , namely , @xmath59 . + on the other hand , the conditions for the onset of peaks of resonance in the linear conductance of the ring connected to the leads , can also be analyzed from the poles of the green functions of the disconnected device .
we state that the poles of @xmath60 will show finite ( or even perfect ) transmission , while those poles of @xmath61 or @xmath62 ( which are not poles of @xmath60 ) will show transmission zeros ( antiresonances ) .
firstly , consider a non degenerate energy eigenvalue @xmath63 of the disconnected device with eigenfunction @xmath64 which is a linear combination of the site orbitals @xmath65 , @xmath66 , with @xmath67 .
if the state @xmath64 has a non vanishing weight at the connection sites 1 and @xmath15 simultaneously , i.e. @xmath68 , then the spectral representation @xmath69 shows that @xmath63 is a pole of @xmath60 because the term @xmath70 is present in the expansion ( [ spectral representation ] ) .
furthermore also the terms @xmath71 will be present in the spectral representation of @xmath61 and @xmath62 , respectively , and @xmath63 will also be a pole of them .
that is , the poles of @xmath60 also become poles of @xmath61 and @xmath62 and all three green functions @xmath26 diverges as @xmath72 , where @xmath73 ( @xmath74 ) is the residue of @xmath26 at the simple pole @xmath63 .
therefore , the green function of the connected ring , eq.([connected g1n ] ) , can be approximated as @xmath75 taking into account that @xmath76 , it reduces to @xmath77 which shows that the transmission has a pole at @xmath78 and a finite transmission @xmath79 .
the pole @xmath63 acquires a finite width proportional to the coupling to the leads @xmath80 . in the particular case
where the sites 1 and @xmath15 are topologically equivalent because of the symmetry of the system , @xmath81 so that perfect transmission occurs .
on the other hand , if @xmath82 is a pole of @xmath61 or @xmath62 , but not of @xmath60 , the numerator @xmath83 of eq .
( [ connected g1n ] ) is finite whilst its denominator diverges ; therefore @xmath84 will show an antiresonance at @xmath82 .
in other words , a finite transmission occurs when the eigenstate @xmath85 of the isolated system have non - vanishing projection on the orbitals @xmath86 and @xmath87 : @xmath88 and @xmath89 .
reciprocally , if one of them equals zero , the electron of energy @xmath82 has a vanishing probability of being at both sites , and therefore no transmission can occur .
+ the case when the eigenvalue @xmath63 is degenerate requires a modification of the above argument .
in such a case , there are more than one states @xmath90 having the same energy @xmath63 . for the sake of simplicity ,
consider just two degenerate eigenfunctions @xmath91 , ( @xmath92 ) .
the spectral representation of the green function now reads @xmath93 the property @xmath94 , valid for the nondegenerate case , does no longer hold here .
instead @xmath95 which can vanish or not depending on the @xmath96 and @xmath97 . if @xmath98 , all above discussion holds ; if not , the real part of the denominator in eq .
( [ green near poles ] ) diverges near @xmath82 as @xmath99 and the transmission becomes suppressed ( @xmath100 ) .
firstly , consider a ring threaded by a magnetic flux with all interdot hopping parameters equal to each other ( @xmath101 ) and interarm coupling @xmath10 .
the on - site energies are taken as @xmath102 , @xmath103 and @xmath104 .
all the results are obtained with a coupling to the leads @xmath105 .
figure [ t(e ) symmetric ring ] shows the dimensionless conductance @xmath106 for the ring connected to the leads in the configuration ( 1,3 ) ( solid line ) , and for the three - sites chains forming the upper ( dashed line ) and lower ( dotted line ) arms with the side - dot 4 and 2 , respectively , coupled by @xmath10 .
the transmission with an applied magnetic flux @xmath107 is shown in dot - dashed line .
transmission function of a ring with interarm coupling in the connection ( 1,3 ) ( solid line ) , and three - sites chains with a laterally coupled side dot representing the upper ( dashed line ) and lower ( dotted line ) arms as a function of the energy @xmath108 .
all dots are connected by the same interdot hopping @xmath109 and interarm couplings @xmath110 , without magnetic flux ( solid line ) and with a flux @xmath107 ( dot - dashed line ) .
( a ) ring with coupling @xmath111 , ( b ) ring with disconnected arms ( @xmath112 ) , and ( c ) ring with coupling @xmath113 .
the @xmath10 couplings were chosen to show the tuning of the antiresonance with the one for the single chain with a lateral dot .
the insets shows , in more detail , the suppression of the antiresonance due to the applied magnetic flux.,width=302 ] the transmission show four peaks at the energy eigenvalues of the systems .
the ring with disconnected arms figure [ t(e ) symmetric ring](b ) shows also an antiresonance , not present in the chains because for @xmath112 there is no side - coupled dot .
this suppression of the transmission in the ring is due to the cancellation of the contributions to the self - energy throughout the upper and lower paths ( @xmath114 ) . in general , eqs .
( [ self - energy contributions 13 ] ) show that the self - energy vanishes if @xmath115=0\end{aligned}\ ] ] in absence of magnetic flux ( @xmath116 ) , @xmath117 vanishes at the energy @xmath118 .
figures [ t(e ) symmetric ring](a)-[t(e ) symmetric ring](c ) depicts the tuning of the antiresonance with @xmath10 , for @xmath119 , @xmath112 and @xmath120 , respectively , such that the antiresonance of the ring can be made to coincide with that from the upper and lower arms with a lateral dot , respectively .
when there is a finite magnetic flux , @xmath117 is complex and its cancellation requires vanishing its real and imaginary parts , i.e , @xmath121 , and @xmath122 .
both equations can not be satisfied simultaneously , except when @xmath123 , which presents an antiresonance at @xmath124 .
therefore , the magnetic field eliminates the antiresonance for a ring with different site energies @xmath125 for arbitrary @xmath10 .
this suppression of the antiresonance is shown in the insets of figures [ t(e ) symmetric ring](a)-[t(e ) symmetric ring](c ) for a flux @xmath107 .
application of gate potentials at dots 2 and 4 allows to tune their on - site energies .
figure [ t12(e , e_dot ) symmetric hopping ] shows the effect on the transmission of varying these parameters and the electron energy @xmath108 with and without magnetic flux for the connection ( 1,2 ) .
bright lines and dark regions represent zones of high and low transmission , respectively .
variation of @xmath126 ( left panels ) was done at @xmath104 while variation of @xmath127 ( right panels ) was done at @xmath103 . in absence of magnetic field ( upper panels ) three peaks of conductance are visible depicted by the bright curves .
the central peak has a linear dependence on the energies @xmath126 and @xmath127 while the external ones are weakly dependent on them , as seen from the slope of the curves .
two antiresonances are also visible , namely , a faint vertical dark line at @xmath128 independent on @xmath7 , and a vertical dark thicker line at @xmath129 ( upper left panel ) and the straight line @xmath129 ( upper right panel ) . hence ,
both @xmath126 and @xmath127 are equally suitable for tuning the maxima of transmission , but only the site 4 , which is not connected to the leads is efficient for tuning the antiresonances .
the lower panels of figure [ t12(e , e_dot ) symmetric hopping ] show the effect of switching on a magnetic flux .
as discussed for the connection ( 1,3 ) , the antiresonances are cancelled out . in particular
, it should be noted that the antiresonance at @xmath128 turned into a peak of transmission .
also the second antiresonance at @xmath129 becomes weakened and the dark sharp straight line turns into a diffuse dark region of low transmission . the logarithm of the transmission function of a ring with coupling between the arms in the connection ( 1,2 ) as a function of the energy @xmath108 and the dots energies @xmath126 and @xmath127 with and without magnetic flux .
all dots are connected by the same interdot hopping parameters @xmath130 , and the magnetic flux was taken as @xmath107 .
bright lines and dark regions represent zones of high and low transmission , respectively .
variation of @xmath126 ( left panels ) was done at @xmath104 while variation of @xmath127 ( right panels ) was done at @xmath103 . ]
figure [ t(e ) fano ] shows in solid lines the transmission through an asymmetric ring where the upper arm has hoppings much smaller than those of the lower arm ( @xmath131 ) .
the dotted blue and dashed red lines are the transmissions of the upper and lower arms separately , calculated as three - site chains with a central site having energy @xmath126 and @xmath127 , respectively , and lateral sites having on site energies @xmath132 .
transmission for the connection ( 1,3 ) as a function of the fermi energy for an asymmetric ring with hopping parameters @xmath133 ( upper arm ) , @xmath134 ( lower arm ) and on - site energies @xmath103 and @xmath135 ( solid line ) , as compared to the transmissions through the upper arm only ( dotted blue line ) , and throughout the lower arm only ( dashed red line ) .
figures ( a)-(c ) correspond to a decoupled ring ( @xmath112 ) ( a ) at zero magnetic field , ( b ) at @xmath136 , and ( c ) magnetic flux @xmath137 .
figure ( d ) has an interarm coupling @xmath138 and magnetic flux @xmath137 .
the insets show in more detail the behaviour of the curves around the resonance . ] in figure [ t(e ) fano ] , the four peaks of conductance , corresponding to connection ( 1,3 ) , can be recognized as those from the lower arm along with the on - site energy of site 2 .
the transmission of the ring nearly coincides in almost the whole range with that of the lower arm , thus showing that the conduction is throughout such a pathway at almost every energy , except for @xmath139 .
when the incident electron is resonant with the site 2 , the transmission is well described by the resonant peak of the upper pathway .
nevertheless , none of the paths by themselves can provide the onset of the antiresonance close to @xmath139 .
the self energy @xmath140 because @xmath141 , except for @xmath142 when they can become comparable .
hence , in a neighbourhood of @xmath142 , the self energy can be approximated as @xmath143 which vanishes for @xmath144 , that is , slightly to the right of the peak @xmath145 , thus giving the fano - like profile . +
the fano - like peak shows the signature of the interference between both paths , typical when a localized state interferes with a continuum .
figures [ t(e ) fano](a)-[t(e ) fano](c ) show the dependence of the fano profile with the magnetic flux . with no magnetic flux , there is the above discussed fano resonance due to the path interference ; the application of a flux @xmath146 suppress the antiresonance leaving only a dip in the transmission which also disappears at @xmath136 leaving only the resonant peak , as seen in [ t(e ) fano](b ) .
further increase of the flux in the range @xmath147 produce a new dip at an energy slightly smaller than @xmath126 while moves the peak to energies slightly higher . at @xmath137
the dip in the transmission of the ring becomes an antiresonance , with the resonant peak tuned with that of the upper chain as seen in the inset of figure [ t(e ) fano](c ) . between @xmath148 and @xmath149 ,
the behaviour of @xmath106 is reversed , such that a cycle is completed in a period of @xmath149 .
figure [ t(e ) fano](d ) depicts the transmission through the ring with an interarm coupling @xmath138 . now
the transmission through the lower ( dashed line ) and upper ( dotted line ) arms , including the site laterally coupled by @xmath10 , shows fano - like resonances at @xmath150 and at @xmath151 , respectively .
the transmission through the ring ( solid line ) still remains close to that of the lower arm with a lateral connection to site 2 .
then , the overall picture for the transmission through a ring with different connection strengths along each arm , is that of the transmission throughout the stronger pathway ( i.e. , the one with larger hoppings ) at almost every energy , except at the one resonant with the energy of the site connecting the arms where a fano interference occurs . in the presence of a magnetic flux , the self energy is a complex quantity and its modulus should be considered .
( [ self - energy contributions 13 ] ) shows that the flux @xmath152 introduces a phase @xmath153 in the paths throughout the arms while no phase change occurs in the self - energy corresponding to the interarm coupling .
let us call @xmath154 the self energy with magnetic field .
then , @xmath155 , where @xmath156 are the real self - energies at zero magnetic field , such that @xmath157 where the first three terms represent the non - interfering transmission along the paths a , b and c. the last two terms contain the effect of the interference due to the quantum and magnetic phases .
it is clearly noted that even for @xmath116 there is an interference between the path contributions to the self - energy .
interestingly , there are two periods in the magnetic phase ; a period @xmath158 ( associated to @xmath159 ) and a period @xmath160 ( associated to @xmath161 ) .
when there is no interarm coupling , @xmath162 , the latter is not present . on the other hand , as soon as a finite @xmath10 exists , the self - energy acquires the longer period modulated by the shorter one .
such a behaviour has been observed in experiments @xcite and were termed as fano resonances of the _ big _ and _ small orbits_. figure [ t(e , phi ) ] shows the transmission ( in a log scale ) @xmath163 as a function of the energy and the magnetic flux , for various values of the interarm coupling @xmath164 0.5 and 1 , and for the two ways of connecting to the leads .
the top ( left and right ) panels , corresponding to @xmath112 , are the only ones showing a period @xmath149 in the flux . as @xmath10 increases , the period @xmath165 becomes apparent .
finally , the bottom ( left and right ) panels show the transmission for a single subring obtained from decoupling the sites 2 and 3 ( i.e , @xmath166 ) , as also done in the experiments @xcite , where the period of the small orbit is clearly apparent .
the logarithm of the transmission coefficient @xmath167 as a function of the electron energy @xmath108 ( horizontal axis ) and the magnetic flux @xmath152 ( vertical axis ) given in units of the quantum of flux @xmath149 , for @xmath112 , 0.5 , and 1 .
all hopping parameters are @xmath168 and all site energies @xmath132 .
the case @xmath112 corresponds to that of a single loop ( _ big orbit _ ) enclosing the magnetic flux .
the bottom pictures correspond to a single subring with @xmath169 and @xmath166 ( _ small orbit _ ) ] in the configuration ( 1,2 ) there are three interfering paths , namely , the direct path through sites 1 and 2 ( @xmath170 ) , the path @xmath171 ( @xmath172 ) and the path @xmath173 ( @xmath174 ) , contributing to the self - energy @xmath175 .
the absolute square of the self - energy turns out @xmath176 where the phase of the big orbit ( @xmath177 ) characterize the interference between the pathways along the arms of the ring , whilst the phase of the small orbit ( @xmath178 ) corresponds to the interference between them and the molecular bond .
the bottom panels of figure [ t(e , phi ) ] , shows the conductance when the hopping @xmath179 has been set equal to zero , such that there is a single small orbit .
nevertheless , it should be noted a few differences visible as dark spots ; they correspond to antiresonances determined by the different topology of the connections .
we have shown the conditions for the onset of resonances and antiresonances in an artificial quantum dot molecule embedded in the arms of an aharonov - bohm interferometer , and its dependence on the tunable parameters .
in general , the peaks of conductance are located at the energy eigenvalue of the isolated device .
a partitioning technique enables to decompose the transmission as a sum along interfering pathways .
the tunability of the coupling between the arms of the interferometer , allows one to weaken or enhance the contribution through the bond of the artificial molecule .
the experimentally observed change in the period of the aharonov - bohm phase from one to twice the quantum of flux is interpreted as due to the opening of one transmission pathway .
our results suggest that a modified coherent single - mode picture including the electron reflection in the subrings forming the small orbits , could also be of help in interpreting the experiments .
application of a magnetic flux leads to a suppression of the antiresonances due to the partial cancellation of the destructive quantum interference .
we have also discussed the differences with a connection not realized experimentally in which one of the dots of the molecule is attached to one lead .
such a configuration could provide other alternatives for tuning the conductance .
this work was partly supported by sgcyt ( universidad nacional del nordeste ) , national agency anpcyt and conicet ( argentina ) under grants pi 112/07 , picto - unne 204/07 and pip 11220090100654/2010 .
99 y. imry , _ introduction to mesoscopic physics _ , 2nd ed .
oxford university press , oxford ( 2002 ) .
m. di ventra , _ electrical transport in nanoscale systems _ , cambridge university press , cambridge ( 2008 ) .
a. g. aronov and yu .
v. shavin rev .
phys . * 59 * , 755 ( 1987 ) .
o. hod , r. baer and e. rabani , j. phys . : condens .
matter * 20 * , 383201 ( 2008 ) .
k. barnham and d. vvedensky ( eds . ) , _ low - dimensional semiconductor structures . fundamentals and device applications _ , cambridge university press , cambridge ( 2001 ) .
y. aharonov and d. bohm , phys . rev . * 115 * , 485 ( 1959 ) .
g. hackenbroich , phys . rep . * 343 * , 463 ( 2001 ) .
o. hod , r. baer and e. rabani , j. phys . chem .
b * 108 * , 14807 ( 2004 ) .
o. hod , r. baer and e. rabani , phys .
* 97 * , 266803 ( 2006 ) .
m. l. ladrn de guevara , f. claro , and p. a. orellana , phys .
rev . * 67 * , 195335 ( 2003 ) .
m. l. ladrn de guevara and p. a. orellana , phys .
73 * , 205303 ( 2006 ) .
i. gmez , f. domnguez - adame and p. orellana , j. phys . : condens .
matter * 16 * , 1613 ( 2004 ) .
e. r. hedin , y. s. joe and a. m. satanin , j. phys . : condens .
matter * 21 * 015303 ( 2009 ) .
an and j .- j .
liu , appl .
. lett . * 96 * , 223508 ( 2010 ) . m. l. ladron de guevara , g. a. lara , p. a. orellana , physica e * 42 * , 1637 ( 2010 ) .
u. fano , phys .
bf 124 , 1866 ( 1961 ) .
k. kobayashi , h. aikawa , s. katsumoto , and y. iye , phys .
lett . * 88 * , 256806 ( 2002 ) .
a. fuhrer , p. brusheim , t. ihn , m. sigrist , k. ensslin , w. wegscheider , and m. bichler , phys .
b * 73 * , 205326 ( 2006 ) .
t. ihn , m. sigrist , k. ensslin , w. wegscheider and m reinwald , new j. phys .
* 9 * , 111 ( 2007 ) .
a. e. miroshnichenko , s. flach and y. s. kivshar , rev .
phys . * 82 * , 2257 ( 2010 ) and references therein .
lwdin , j. math . phys . * 3 * , 969 ( 1962 ) .
t. hansen , g. c. solomon , d. q. andrews and m. a. ratner , j. chem . phys . * 131 * , 194704 ( 2009 ) . | we study theoretically the transmission through a quantum dot molecule embedded in the arms of an aharonov - bohm four quantum dot ring threaded by a magnetic flux .
the tunable molecular coupling provides a transmission pathway between the interferometer arms in addition to those along the arms . from a decomposition of the transmission in terms of contributions from paths
, we show that antiresonances in the transmission arise from the interference of the self - energy along different paths and that application of a magnetic flux can produce the suppression of such antiresonances .
the occurrence of a period of twice the quantum of flux arises to the opening of transmission pathway through the dot molecule .
two different connections of the device to the leads are considered and their spectra of conductance are compared as a function of the tunable parameters of the model . |
training neural networks requires setting several hyper - parameters to adequate values : number of hidden layers , number of neurons per hidden layer , learning rate , momentum , dropout rate , etc .
although tuning such hyper - parameters via parameter search has been recently investigated by @xcite , doing so remains extremely costly , which makes it imperative to develop more efficient techniques . of the many hyper - parameters , those that determine the network s architecture are among the hardest to tune , especially because changing them during training is more difficult than adjusting more dynamic parameters such as the learning rate or momentum .
typically , the architecture parameters are set once and for all before training begins .
thus , assigning them correctly is paramount : if the network is too small , it will not learn well ; if it is too large , it may take significantly longer to train while running the risk of overfitting .
networks are therefore usually trained with more parameters than necessary , and pruned once the training is complete .
this paper introduces divnet , a new technique for reducing the size of a network . divnetdecreases the amount of redundancy in a neural network ( and hence its size ) in two steps : first
, it samples a diverse subset of neurons ; then , it merges the remaining neurons with the ones previously selected .
specifically , divnetmodels neuronal diversity by placing a determinantal point process ( dpp ) @xcite over neurons in a layer , which is then used to select a subset of diverse neurons .
subsequently , divnet``fuses '' information from the dropped neurons into the selected ones through a reweighting procedure .
together , these steps reduce network size ( and act as implicit regularization ) , without requiring any further training or significantly hurting performance .
divnetis fast and runs in time negligible compared to the network s prior training time .
moreover , it is agnostic to other network parameters such as activation functions , number of hidden layers , and learning rates . for simplicity ,
we describe and analyze divnetfor feed - forward neural networks , however divnetis _ not _ limited to this setting . indeed , since divnetoperates on a layer fully connected to the following one in a network s hierarchy
, it applies equally well to other architectures with fully connected layers .
for example , it can be applied without any further modification to deep belief nets and to the fully - connected layers in convolutional neural networks . as these layers are typically responsible for the large majority of the cnns memory footprint @xcite , divnetis particularly well suited for such networks .
[ [ contributions . ] ] contributions .
+ + + + + + + + + + + + + + the key contributions of this paper are the following : =1em introduction of dpps as a flexible , powerful tool for modeling layerwise neuronal diversity ( [ sec : pruning ] ) .
specifically , we present a practical method for creating dpps over neurons , which enables diversity promoting sampling and thereby leads to smaller network sizes .
a simple but crucial `` fusing '' step that minimizes the adverse effects of removing neurons .
specifically , we introduce a reweighting procedure for a neuron s connections that transfers the contributions of the pruned neurons to the ones that are retained ( [ sec : reweighting ] ) .
the combination of both ideas is called divnet .
we perform several experiments to validate divnetand compare to previous neuron pruning approaches , which divnetconsistently outperforms . notably , divnet s reweighting strategy benefits other pruning approaches . [ [ related - work . ] ] related work .
+ + + + + + + + + + + + + due to their large number of parameters , deep neural networks typically have a heavy memory footprint .
moreover , in many deep neural network models parameters show a significant amount of redundancy @xcite .
consequently , there has been significant interest in developing techniques for reducing a network s size without penalizing its performance . a common approach to reducing
the number of parameters is to remove connections between layers . in @xcite
, connections are deleted using information drawn from the hessian of the network s error function . @xcite
reduce the number of parameters by analyzing the weight matrices , and applying low - rank factorization to the final weight layer .
@xcite remove connections with weights smaller than a given threshold before retraining the network .
these methods focus on deleting parameters whose removal influences the network the least , while divnetseeks diversity and merges similar neurons ; these methods can thus be used in conjunction with ours .
although methods such as @xcite that remove connections between layers may also delete neurons from the network by removing all of their outgoing or incoming connections , it is likely that the overall impact on the size of the network will be lesser than approaches such as divnetthat remove entire neurons : indeed , removing a neuron decreases the number of rows or columns of the weight matrices connecting the neuron s layer to both the previous and following layers .
convolutional neural networks @xcite replace fully - connected layers with convolution and subsampling layers , which significantly decreases the number of parameters .
however , as cnns still maintain fully - connected layers , they also benefit from divnet .
closer to our approach of reducing the number of hidden neurons is @xcite , which evaluates each hidden neuron s importance and deletes neurons with the smaller importance scores . in @xcite
, a neuron is pruned when its weights are similar to those of another neuron in its layer , leading to a weight update procedure that is somewhat similar in idea ( albeit simpler ) to our reweighting step : where @xcite removes neurons with equal or similar weights , we consider the more complicated task of merging neurons that , as a group , perform redundant calculations based on their activations .
other recent approaches consider network compression without pruning : in @xcite , a new , smaller network is trained on the outputs of the large network ; @xcite use hashing to reduce the size of the weight matrices by forcing all connections within the same hash bucket to have the same weight . @xcite and @xcite
show that networks can be trained and run using limited precision values to store the network parameters , thus reducing the overall memory footprint .
we emphasize that divnet s focus on neuronal diversity is orthogonal and complementary to prior network compression techniques .
consequently , divnetcan be combined , in most cases trivially , with previous approaches to memory footprint reduction .
in this section we introduce our technique for modeling neuronal diversity more formally .
let @xmath0 denote the training data , @xmath1 a layer of @xmath2 neurons , @xmath3 the activation of the @xmath4-th neuron on input @xmath5 , and @xmath6 the activation vector of the @xmath4-th neuron obtained by feeding the training data through the network . to enforce diversity in layer @xmath1 ,
we must determine which neurons are computing redundant information and remove them . doing so requires finding a maximal subset of ( linearly ) independent activation vectors in a layer and retaining only the corresponding neurons . in practice , however , the number of items in the training set ( or the number of batches ) can be much larger than the number of neurons in a layer , so the activation vectors @xmath7 are likely linearly independent .
merely selecting by the maximal subset may thus lead to a trivial solution that selects all neurons . reducing redundancy
therefore requires a more careful approach to sampling .
we propose to select a subset of neurons whose activation patterns are diverse while contributing to the network s overall computation ( i.e. , their activations are not saturated at 0 ) .
we achieve this diverse selection by formulating the neuron selection task as sampling from a determinantal point process ( dpp ) .
we describe the details below .
dpps are probability measures over subsets of a ground set of items .
originally introduced to model the repulsive behavior of fermions @xcite , they have since been used fruitfully in machine - learning @xcite .
interestingly , they have also been recently applied to modeling inter - neuron inhibitions in neural spiking behavior in the rat hippocampus @xcite .
dpps present an elegant mathematical technique to model diversity : the probability mass associated to each subset is proportional to the determinant ( hence the name ) of a dpp kernel matrix .
the determinant encodes negative associations between variables , and thus dpps tend to assign higher probability mass to diverse subsets ( corresponding to diverse submatrices of the dpp kernel ) .
formally , a ground set of @xmath8 items @xmath9 and a probability @xmath10 $ ] such that @xmath11 where @xmath12 is a @xmath8-by-@xmath8 positive definite matrix , form a dpp .
@xmath12 is called the _ dpp kernel _ ; here , @xmath13 indicates the @xmath14 principal submatrix of @xmath12 indexed by the elements of @xmath15 .
the key ingredient that remains to be specified is the dpp kernel , which we now describe .
there are numerous potential choices for the dpp kernel .
we found that experimentally a well - tuned gaussian rbf kernel provided a good balance between simplicity and quality : for instance , it provides much better results that simple linear kernels ( obtained via the outer product of the activation vectors ) and is easier to use than more complex gaussian rbf kernels with additional parameters .
a more thorough evaluation of kernel choice is future work .
recall that layer @xmath1 has activations @xmath7 . using these , we first create an @xmath16 kernel @xmath17 with bandwidth parameter @xmath18 by setting @xmath19 to ensure strict positive definiteness of the kernel matrix @xmath17 , we add a small diagonal perturbation @xmath20 to @xmath17 ( @xmath21 ) .
the choice of the bandwidth parameter could be done by cross - validation , but that would greatly increase the training cost .
therefore , we use the fixed choice @xmath22 , which was experimentally seen to work well . finally , in order to limit rounding errors , we introduce a final scaling operation : suppose we wish to obtain a desired size , say @xmath23 , of sampled subsets ( in which case we are said to be using a @xmath23-dpp @xcite ) . to that end , we can scale the kernel @xmath24 by a factor @xmath25 , so that its _ expected _ sample size becomes @xmath23 . for a dpp with kernel @xmath12
, the expected sample size is given by ( * ? ? ?
34 ) : @xmath26 = \operatorname{tr}(l ( i+l)^{-1}).\ ] ] therefore , we scale the kernel to @xmath27 with @xmath25 such that @xmath28 where @xmath29 is the expected sample size for the kernel @xmath24 . finally , generating and then sampling from @xmath30 has @xmath31 cost . in our experiments , this sampling cost was negligible compared with the cost of training . for networks with very large hidden layers
, one can avoiding the @xmath32 cost by using more scalable sampling techniques @xcite . simply excising the neurons that are not sampled by the dpp
drastically alters the neuron inputs to the next layer .
intuitively , since activations of neurons marked redundant are not arbitrary , throwing them away is wasteful .
ideally we should preserve the total information of a given layer , which suggests that we should `` fuse '' the information from unselected neurons into the selected ones .
we achieve this via a reweighting procedure as outlined below . without loss of generality , let neurons 1 through @xmath23 be the ones sampled by the dpp and @xmath33 their corresponding activation vectors .
let @xmath34 be the weights connecting the @xmath4-th neuron ( @xmath35 ) in the current layer to the @xmath36-th neuron in the next layer ; let @xmath37 denote the updated weights after merging the contributions from the removed neurons .
we seek to minimize the impact of removing @xmath38 neurons from layer @xmath1 . to that end
, we minimize the difference in inputs to neurons in the subsequent layer before ( @xmath39 ) and after ( @xmath40 ) dpp pruning .
that is , we wish to solve for all neurons in the next layer ( indexed by @xmath36 , @xmath41 ) : @xmath42 eq . [ eq : reweighting ] is minimized when @xmath43 is the projection of @xmath44 onto the linear space generated by @xmath45 .
thus , to minimize eq .
[ eq : reweighting ] , we obtain the coefficients @xmath46 that for @xmath47 minimize @xmath48 and then update the weights by setting @xmath49 using ordinary least squares to obtain @xmath50 , the reweighting procedure runs in @xmath51 .
to quantify the performance of our algorithm , we present below the results of experiments on common datasets for neural network evaluation : ` mnist ` @xcite , ` mnist_rot ` @xcite and ` cifar-10 ` @xcite .
all networks were trained up until a certain training error threshold , using softmax activation on the output layer and sigmoids on other layers ; see table [ tab : networks ] for more details . in all following plots ,
error bars represent standard deviations .
.overview of the sets of networks used in the experiments .
we train each class of networks until the first iteration of backprop for which the training error reaches a predefined threshold . [ cols="^,^,^,^",options="header " , ] to validate our claims on the benefits of using dpps and fusing neurons , we compare these steps separately and also simultaneously against random pruning , where a fixed number of neurons are chosen uniformly at random from a layer and then removed , with and without our fusing step .
we present performance results on the test data ; of course , both dpp selection and reweighting are based solely on information drawn from the training data .
figure [ fig : reconstruction ] visualizes neuron activations in the first hidden layer of a network trained on the ` mnist ` dataset .
each column in the plotted heat maps represents the activation of a neuron on instances of digits 0 through 9 .
figure [ fig : reconstruction - dpp ] shows the activations of the 50 neurons sampled using a @xmath23-dpp ( @xmath23 = 50 ) defined over the first hidden layer , whereas figure [ fig : reconstruction - first ] shows the activations of the first 50 neurons of the same layer .
figure [ fig : reconstruction - first ] contains multiple similar columns : for example , there are 3 entirely green columns , corresponding to three neurons that saturate to 1 on each of the 10 instances . in contrast , the dpp samples neurons with diverse activations , and figure [ fig : reconstruction - dpp ] shows no similar redundancy
. figures [ fig : k - dpp ] through [ fig : reweighting - pruning - influence ] illustrate the impact of each step of divnetseparately .
figure [ fig : k - dpp ] shows the impact of pruning on test error using dpp pruning and random pruning ( in which a fixed number of neurons are selected uniformly at random and removed from the network ) .
dpp - pruned networks have consistently better training and test errors than networks pruned at random for the same final size .
as expected , the more neurons are maintained , the less the error suffers from the pruning procedure ; however , the pruning is in both cases destructive , and is seen to significantly increase the error rate .
this phenomenon can be mitigated by our reweighting procedure , as shown in figure [ fig : reweighting - influence ] . by fusing and reweighting neurons after pruning , the training and test errors
are considerably reduced , even when 90% of the layer s neurons are removed .
pruning also reduces variability of the results : the standard deviation for the results of the reweighted networks is significantly smaller than for the non - reweighted networks , and may be thus seen as a way to regularize neural networks .
finally , figure [ fig : reweighting - pruning - influence ] illustrates the performance of divnet(dpp pruning and reweighting ) compared to random pruning with reweighting .
although divnet s performance is ultimately better , the reweighting procedure also dramatically benefits the networks that were pruned randomly .
.45 [ fig : k - dpp - test - error ] -.4 cm .45 [ fig : k - dpp - training - error ] -.4 cm .45 [ fig : reweighting - influence - test - error ] -.4 cm .45 [ fig : reweighting - influence - training - error ] -.4 cm .45 [ fig : reweighting - pruning - influence - test - error ] -.4 cm .45 [ fig : reweighting - pruning - influence - training - error ] -.4 cm we also ran these experiments on networks for shrinking the second layer while maintaining the first layer intact .
the results are similar , and may be found in appendix [ appendix : second - layer ] .
notably , we found that the gap between divnetand random pruning s performances was much wider when pruning the last layer .
we believe this is due to the connections to the output layer being learned much faster , thus letting a small , diverse subset of neurons ( hence well suited to dpp sampling ) in the last hidden layer take over the majority of the computational effort .
much attention has been given to reducing the size of neural networks in order to reduce memory consumption .
when using neural nets locally on devices with limited memory , it is crucial that their memory footprint be as small as possible . .45 [ fig : activation - pruning - mnist ] -.4 cm .45 [ fig : activation - pruning - rot ] -.4 cm node importance - based pruning ( henceforth `` importance pruning '' ) is one of the most intuitive ways to cut down on network size
. introduced to deep networks by @xcite , this method removes the neurons whose calculations impact the network the least . among the three solutions to estimating a neuron s importance discussed in @xcite , the sum the output weights of each neuron ( the ` onorm ' function ) provided the best results : @xmath52 figure [ fig : activation - pruning ]
compares the test data error of the networks after being pruned using importance pruning that uses onorm as a measure of relevance against divnet .
since importance pruning deletes neurons that contribute the least to the next layer s computations , it performs well up to a certain point ; however , when pruning a significant amount of neurons , this pruning procedure even removes neurons performing essential calculations , hurting the network s performance significantly .
however , since divnetfuses redundant neurons , instead of merely deleting them its resulting network delivers much better performance even with very large amounts of pruning .
in order to illustrate numerically the impact of divneton network performance , table [ tab : compression - train ] shows network training and test errors ( between 0 and 1 ) under various compression rates obtained with divnet , without additional retraining ( that is , the pruned network is not retrained to re - optimize its weights ) .
@xmath53=1em in all experiments , sampling and reweighting ran several orders of magnitude faster than training ; reweighting required significantly more time than sampling .
if divnetmust be further sped up , a fraction of the training set can be used instead of the entire set , at the possible cost of subsequent network performance .
when using dpps with a gaussian rbf kernel , sampled neurons need not have linearly independent activation vectors : not only is the dpp sampling probabilistic , the kernel itself is not scale invariant . indeed , for two collinear but unequal activation vectors , the corresponding coefficient in the kernel will not be 1 ( or @xmath25 with the @xmath54 update ) .
in our work , we selected a subset of neurons by sampling once from the dpp .
alternatively , one could sample a fixed amount of times , using the subset with the highest likelihood ( i.e. , largest @xmath55 ) , or greedily approximate the mode of the dpp distribution . our reweighting procedure benefits competing pruning methods as well ( see figure [ fig : reweighting - pruning - influence ] ) .
we also investigated dpp sampling for pruning concurrently with training iterations , hoping that this might allow us to detect superfluous neurons before convergence , and thus reduce training time .
however , we observed that in this case dpp pruning , with or without reweighting , did not offer a significant advantage over random pruning .
consistently over all datasets and networks , the expected sample size from the kernel @xmath17 was much smaller for the last hidden layer than for other layers .
we hypothesize that this is caused by the connections to the output layer being learned faster than the others , allowing a small subset of neurons to take over the majority of the computational effort .
divnetleverages similarities between the behaviors of neurons in a layer to detect redundant parameters and merge them , thereby enforcing neuronal diversity within each hidden layer . using divnet , large
, redundant networks can be shrunk to much smaller structures without impacting their performance and without requiring further training .
we believe that the performance profile of divnetwill remain similar even when scaling to larger scale datasets and networks , and hope to include results on bigger problems ( e.g. , imagenet @xcite ) in the future .
many hyper - parameters can be tuned by a user as per need include : the number of remaining neurons per layer can be fixed manually ; the precision of the reweighting and the sampling procedure can be tuned by choosing how many training instances are used to generate the dpp kernel and the reweighting coefficients , creating a trade - off between accuracy , memory management , and computational time .
although divnetrequires the user to select the size of the final network , we believe that a method where no parameter explicitly needs to be tuned is worth investigating .
the fact that dpps can be augmented to also reflect different distributions over the sampled set sizes @xcite might be leveraged to remove the burden of choosing the layer s size from the user .
importantly , divnetis agnostic to most parameters of the network , as it only requires knowledge of the activation vectors .
consequently , divnetcan be easily used jointly with other pruning / memory management methods to reduce size .
further , the reweighting procedure is agnostic to how the pruning is done , as shown in our experiments . furthermore , the principles behind divnetcan theoretically also be leveraged in non fully - connected settings .
for example , the same diversifying approach may also be applicable to filters in cnns : if a layer of the cnn is connected to a simple , feed - forward layer such as the s4 layer in @xcite by viewing each filter s activation values as an vector and applying divneton the resulting activation matrix , one may be able to remove entire filters from the network , thus significantly reducing cnn s memory footprint .
finally , we believe that investigating dpp pruning with different kernels , such as kernels invariant to the scaling of the activation vectors , or even kernels learned from data , may provide insight into which interactions between neurons of a layer contain the information necessary for obtaining good representations and accurate classification .
this marks an interesting line of future investigation , both for training and representation learning .
this work is partly supported by nsf grant : iis-1409802 .
.49 .49 .49 .49 .49 .49
on training error ( using the networks trained on ` mnist ` ) .
the dotted lines show min and max errors . ] on the number of neurons that remain after pruning networks trained on ` mnist ` ( when pruning non - parametrically , using a dpp instead of a @xmath23-dpp . ) ] | we introduce divnet , a flexible technique for learning networks with diverse neurons .
divnetmodels neuronal diversity by placing a determinantal point process ( dpp ) over neurons in a given layer .
it uses this dpp to select a subset of diverse neurons and subsequently fuses the redundant neurons into the selected ones .
compared with previous approaches , divnetoffers a more principled , flexible technique for capturing neuronal diversity and thus implicitly enforcing regularization .
this enables effective auto - tuning of network architecture and leads to smaller network sizes without hurting performance .
moreover , through its focus on diversity and neuron fusing , divnetremains compatible with other procedures that seek to reduce memory footprints of networks .
we present experimental results to corroborate our claims : for pruning neural networks , divnetis seen to be notably superior to competing approaches . |
fig.1 @xmath91 ( see text ) as a function of @xmath134 for two different spectral functions as shown in the inset . | the interrelation between the condensation energy and the optical sum rules has been investigated .
it has been shown that the so called partial sum rule violation is related mainly to a temperature dependence of the relaxation rate rather than to the appearance of superconductivity itself .
moreover , we demonstrate that the experimental data on the temperature dependence of the optical sum rule can be explained rather well by an account of strong electron - phonon interaction .
many recently published works are concerned with the origin of the condensation energy of the superconducting state , a possible violation of so - called optical sum rules , and the relation between these phenomena .
these papers include both theoretical investigationshm , nor1,scal , legg , nor2,kim of these problems and experimental attemptsbasov , vdm , holcomb , santan to observe a violation of the optical sum rule .
usually , the possibility of such violation is related to the change of the kinetic energy of metals under superconducting transition . if this statement would be correct , than the violation of the optical sum rule should be most clear seen in the bardeen - cooper - schrieffer@xcite@xmath0 type superconductors .
it has been proved exactly by bogolyubov@xcite that the original model does not contain any potential energy .
it is easy to see considering the original @xmath1 hamiltonian @xmath2n .
bogolyubov has proven that for the normal state @xmath3where @xmath4 is the system volume .
it means that in the thermodynamic limit @xmath5the @xmath1 hamiltonian can be exactly diagonalized in the superconducting state using the bogolyubov - valatin transformation @xmath6it leads to @xmath7where @xmath8which is a hamiltonian of noninteracting but superconducting quasiparticles .
the condensation energy arises from the decreasing of the ground state eigenvalue of the expression @xmath9 due to an appearance of a gap @xmath10 .
the same is true for any mechanisms of superconductivity .
the decrease of a properly defined one - quasiparticle energy due to the appearance of the gap on a fermi level is the main contribution to the condensation energy .
this phenomenon has a certain feature is common to a metal - insulator transition , where the band structure contribution to the total energy decreases also due to the appearance of the gap on the fermi level .
the example of the bcs model shows that the division of the total energy into kinetic and potential parts is not a trivial problem even for weakly interacting quasiparticles .
this division becomes even worse defined in systems with strongly interacting electrons .
the study of the optical sum rule and its change below the superconducting transition should be based , from our point of view , on calculations of the conductivity itself and a detailed analysis of this function and its dependence on temperature and frequency .
the optical sum rule can be written in general form as @xmath11where @xmath12 is the real part of the dynamical conductivity , @xmath13 is the total electron density and @xmath14 is the bare electron mass .
the function @xmath12 has a zero frequency @xmath15 function peak in the superconducting state due to the dissipationless ( rigid ) response of the superconducting condensate .
the amplitude of this peak @xmath16 can be expressed in terms of an corresponding penetration depth @xmath17 @xmath18where @xmath19 is the velocity of light .
the existence of this @xmath15 function contribution to @xmath12 in the superconducting state leads to the so - called ferrel - glover - tinkham sum rule@xcite @xmath20 = \frac{c^{2}}{8\lambda _ { l}^{2}% } \ , \label{3}\]]where @xmath21 is the conductivity in the normal and superconducting states , correspondingly . in such general form
this sum rule can never be violated for any superconductors possessing ideal diamagnetic response with a finite penetration depth @xmath17 .
the real measurement of the dynamical conductivity can never be made up to infinite frequency .
they are restricted in practice to some finite value @xmath22 . as it is well known@xcite , the sum rule ( 9 ) is totally satisfied in conventional superconductors when the integration is performed up to @xmath23 where @xmath10 is the superconducting gap .
this value of @xmath24 is of the order of the characteristic phonon energies .
the statements about the sum rules violation , which has been made in the experimental papersbasov , vdm , holcomb , santan , mean that the value @xmath24 in high-@xmath25 superconductors is much larger than in conventional ones .
the maximum value @xmath22 in high-@xmath25 systems , if they were also conventional , should be @xmath26 because they have a magnitude of the gap @xmath27 it has been shown in@xcite that for the interlayer conductivity the optical sum rules are not saturated at least in underdoped regime for @xmath28 even more intriguing results have been obtained recently in the paper@xcite where the violation sum rules have been observed up to very high energies @xmath29 the main goal of the present paper is to show that the observed violation of the optical sum rules at least for @xmath30 is not related explicitly to any mechanism of superconductivity .
this violation is the direct consequence of a high value of the electron relaxation rate @xmath31 , critical temperature @xmath32 and @xmath10 themselves .
let us consider the for the future discussion important the so - called restricted or partial optical sum rule .
usually it is used in the form @xmath33here @xmath34is the an effective inverse electron band mass , which is defined as @xmath35where @xmath36 is an electron distribution function . for a hamiltonian with nearest neighbor hopping @xmath37
can be presented in terms of a average of the one band kinetic energy @xcite @xmath38@xmath39here @xmath40 is the lattice spacing in @xmath41 - direction and @xmath42 is the nearest neighbor hopping integral .
usually it is believed that the high energy cut off frequency @xmath24 should be chosen of the order of the corresponding band plasma frequency @xmath43 and is much smaller than the energies of interband transitions .
this partial sum rule can be easily proved for noninteracting band electrons including the _ interband _ transitions in the expression for the conductivity @xmath44here the summation is over all empty high energy bands with an electron dispersion @xmath45 the minimum of the value @xmath46 is the energy of the interband transitions .
now , using @xmath47 we can easily prove the sum rules @xmath48 and @xmath49 which give the well known identity for the electron inverse effective mass@xcite @xmath50however is it not the case for interacting electrons .
this fact can be easily understood using the model of electrons interacting with impurities .
the _ intraband _ contribution to the optical conductivity can be written in this case in form of the usual drude expression for @xmath51 @xmath52 where @xmath53 is the relaxation rate due to impurity scattering .
one can derive the well known result for the partial optical sum rule @xmath54 this example shows that the _ intraband _ sum rules @xmath48and @xmath49 can be satisfied in the presence of the interaction only in the limit @xmath55 .
it is also true for any interactions other than impurity scattering including , for example , the electron - phonon interaction .
moreover , this violation of the optical sum rules can not be obtained from the calculation of the kinetic energy change from @xmath56 as it was made in the refs.@xcite .
as it follows from @xmath57 the optical sum rules violation depends on the high energy cutoff @xmath24 but this parameter is certainly absent in the expression for the kinetic energy .
the _ interband _ transitions will also be changed due to electron interactions .
we can rewrite @xmath58 for the impurity scattering model in the simplest approximation as @xmath59 the general sum rule @xmath60 is certainly satisfied in this model for any value of the relaxation rate @xmath61 but it is not true for the partial sum rule as we have discussed above .
further , as it follows from @xmath62 the _ interband _ contribution to the conductivity becomes now spread out over all intervals of energies including very low @xmath63 it means that we can not even divide experimental data into _ intraband _ contributions and _ interband _ ones .
all these effects are small as @xmath64 and @xmath65 however , if we shall take into account that the relaxation rate in high-@xmath25 compounds can reach values @xmath66 we see immediately that these effects can be very important even in the normal state . as the discussion given above confirms
, there is no other way to understand the behavior of the partial sum rules than to calculate the conductivity itself as a function of frequency , temperature , doping etc .
it can be obtained by a calculation of a current - current correlation function .
we would like to emphasize here that expression for the current - current correlation function does not contain , at least in the absence of the vertex corrections , any explicit information about the mechanism of superconductivity .
all implicit information about the mechanism is contained in the one - particle green s function which can be written as @xmath67here @xmath68 is a renormalization function , @xmath69 is the superconducting order parameter , and @xmath70 are pauli matrices .
these functions should be calculated in turn from the general equations for the green s function of electrons .
such equations have been derived by eliashberg@xcite for conventional superconductors with the electron - phonon pairing mechanism and , for example , in refs.@xcite for @xmath71wave superconductivity .
the expression for @xmath72 can be written for @xmath73 in the form @xmath74the expression on the right hand side of the @xmath75 is nothing else than the bcs type coherency factors .
the @xmath75 have been derived@xcite for the conventional superconductors and was used recently@xcite to discuss the problem of the influence of the superconducting gap on the optical spectra at @xmath76 . in spite of
that this expression has been derived in the framework of the usual @xmath77wave superconductivity , it has with a slight modification much wider areas of applications .
one can , for example , include the angular dependence of the gap for anisotropic superconductors and perform the integration over the angle .
it is easy to show using @xmath78 that for @xmath79 we have @xmath80 here the numerical coefficient @xmath81 is of the order of unity and it is included to take into account the possible averaging of the angular dependence of the gap function .
the same estimation for the dynamical conductivity of a superconducting state at @xmath82 can be obtained from the equations derived in ref.@xcite .
@xmath83 shows that the direct contribution of the superconducting gap to the dynamical conductivity and , therefore , to the optical sum rules has the same smallness for any mechanism of superconductivity , that is @xmath84 .
this smallness is , certainly , different for conventional superconductors and high-@xmath25 ones because the gap in the later is one order larger .
@xmath85 for the _ interband _ contribution can also be generalized for the superconducting case and it can be shown that their difference from the normal state is of the order of @xmath86 we shall not consider the behavior of the optical sum rules in the superconducting state further in this paper because the real mechanism of superconductivity in high-@xmath25 systems is unknown . the investigation of the optical sum rules for the normal state will be of our main interest in the rest part of this paper .
the detailed experimental study of this problem has been done by d. van der marel and coworkers@xcite .
they have measured the conductivity in a wide frequency range and temperature intervals and than two optical sum rules have been calculated .
one of them was the low energy sum rule @xmath87 @xmath88 where @xmath89 is the penetration depth in the @xmath90 plane and the other one is the high energy sum rule @xmath91 @xmath92 a few very prominent features in the behavior of both @xmath87 and @xmath91have been found in this work .
first , @xmath87 and @xmath91 are temperature dependent in the superconducting state as well as in the normal one .
second , this dependence , at least , in the normal state is well described by a quadratic function of @xmath93 in addition , the low energy part depends on the high - energy cutoff frequency @xmath94 if we consider the cases @xmath95 and @xmath96 we are coming now to the consideration of details in the high energy part of the optical sum rule @xmath97 preliminary we shall neglect the direct contributions of the superconducting gap to the value of @xmath91 because the ratio @xmath98 for the considered values of frequencies is very small @xmath99 .
the main problem in the calculation of the normal state conductivity is to establish the origin of the electron relaxation in high-@xmath25 systems .
this problem along with the problem of the origin of superconductivity itself has been disputed during the last 15 years .
it was shown ( see for details@xcite ) that the main source of the relaxation processes in the normal state of high-@xmath25 superconductors is the strong electron - phonon interaction .
recently , it has been additionally demonstrated through examination of the frequency and temperature dependence of the optical reflectivity in the @xmath100 systemmaksssc that this interaction leads to a strong temperature dependence of the conductivity up to very high frequencies .
the experimental verification of the existence of strong electron - phonon interaction in high-@xmath25 superconductors has been also obtained in arpes measurements@xcite as an effect of an electron mass renormalization .
there is some discussion@xcite about the possibility , that the electron mass renormalization observed in ref.@xciteand the corresponding change of the relaxation rate can been explained by the interaction with the so - called magnetic resonance peak. this possibility , however , is unlikely@xcite , at least , for the normal state . as is well known , in the normal state the conductivity @xmath101 in a presence of the strong electron - phonon interaction can be written in the form@xcite @xmath102 where @xmath103 is the frequency dependent optical mass and @xmath31 is the optical relaxation rate .
the readers can find the precise expressions for both these functions in terms of the eliashberg function in refs.@xcite and we shall not reproduce them here .
@xmath104 for high values of frequencies @xmath105 can be rewritten for the real part of the conductivity in the form @xmath106 where @xmath107 is independent on frequency@xcite @xmath108 here @xmath109 is the transport eliashberg function .
it is easy to show by using @xmath110 , that @xmath111where @xmath112 is the _ transport _ constant of epi , and @xmath113 is the average phonon frequency . at considerably high temperatures , on the other hand , @xmath107 can be written as @xmath114 @xmath115 and @xmath116 show that the relaxation rate can increase considerably with increasing of temperature .
it will lead , to some increase of the high frequency part of the optical sum rule , which can be written in the form @xmath117 where @xmath118 . using @xmath119 and @xmath120
we can easily calculate this value .
there are two independent fitting parameters in this procedure : the plasma frequency @xmath121 and the coupling constant @xmath122 .
we have chosen the value @xmath123 in accordance with apres data @xcite .
the value of the _ intraband _ plasma frequency is also unknown , but it is bounded from above by the value of the low part of the sum rule obtained in ref.@xcite , that is @xmath124 for the numerical calculations we have employed the eliashberg function from our preceding papers@xcite and use the general expression for the conductivity @xmath125 rather than approximate @xmath126 .
we carried out our calculations for two slightly different eliashberg functions shown in the inset in fig.1 , having the same value of @xmath127 the difference between these spectra is related to the different coupling of electrons with a soft phonon @xmath128 and harder ones .
the temperature dependence of @xmath91 is shown in fig.1 at @xmath129 we have used as in ref.@xcite the @xmath130 scale for the temperature to demonstrate the near perfect quadratic dependence @xmath91 on @xmath93 the overall agreement of our results with the experimental data is reasonably well .
the same is true concerning the experimentally observed difference @xmath131it can also be seen from fig.1 , that @xmath91 has a temperature dependence also at @xmath132 where @xmath25 is the critical temperature of the superconducting transition ( @xmath133 for the considered case ) . *
* we would also like to emphasize that there is a little different behavior of these curves at low temperatures .
the value @xmath91 decreases with decreasing of temperature for a softer spectrum even faster for low temperature than @xmath134 as it has for high @xmath135 .
in contrast , @xmath91 has a more weaker temperature dependence at low @xmath135 for more harder spectrum .
the difference @xmath136 for the spectrum with soft low frequency phonons is larger then for the harder one . *
* we did not take into account in our calculations the influence of the superconductivity on @xmath97 it is small from our point of view but it can exist .
it is clear from the above consideration that it is very difficult to separate using experimental data this specific superconducting contribution from the total change of @xmath91 connected with the change of the relaxation rate . *
* the general behavior of the low energy sum rule @xmath137 ( not shown in fig.1 ) is also reproduced rather well in our approach , at least for the normal state .
there is only one contradiction related to the total amplitude of the change of the value @xmath138 it is clear from the above consideration that the following equality should be satisfied in the normal state @xmath139 where @xmath140 it is easy to see that @xmath141 it means that the total change of the low energy sum rule @xmath142 should be twice larger than the change of @xmath143 . measurements@xcite
give rather the value @xmath144 instead of two .
we do not know the exact origin of this contradiction .
it is possible that it is related to the temperature dependence of the _ interband _ transitions which have not been included into our calculations .
indeed , we have obtained as the value of the intraband contribution to @xmath145 at @xmath146 .
it is much smaller than the experimentally measured @xmath91 @xmath147 the difference between these two values comes from the _ interband _ transitions .
it is difficult to say anything definite at this time about the temperature dependence of the _ interband _ transitions and we will continue our activity in this direction .
now we would like to emphasize that it not easy to find any other mechanism of the relaxation besides the electron - phonon one which can lead to a temperature dependence of the relaxation rate at so high frequencies .
many of them , including , for example , the marginal fermi liquid@xcite , do not give any temperature dependence at @xmath148 we can not calculate and compare with the experiment the sum rules in the underdoped regime due to the existence of a _ pseudogap _ phenomenon because its origin is also unknown .
we can , however , claim that it is very likely that the discussed effect will be also exist in the underdoped case .
this statement is based on observations obtained both by optical measurements @xcite and as well as arpes@xcite , that the relaxation rate increases with decreasing the doping level .
this also is confirmed by result obtained in ref.@xcite on the sum rules violation . in summary , we have shown that there are no new energy scales defining the influence of the superconductivity on the intraband contribution to the optical sum rules , besides the superconducting energy gap itself .
this is true for any mechanism of the superconductivity because the expression for @xmath149 at high frequencies does not include any explicit information about such mechanisms . the experimentally observed violation of the sum rule is mainly related to the properties of the normal state of high-@xmath25 superconductors and it is ruled mainly by the frequency and temperature dependence of the relaxation rate . we also have shown that the experimental data obtained in the ref.@xcite can be explained very reasonably in the framework of the usual model with strong electron - phonon interaction .
the consideration of the sum rules for the interplanar conductivity where the coherent transport is absent in the normal state requires a more serious approach and the knowledge of the mechanism blocking this transport .
we should also know more details about the pseudogap phenomenon and its interplay with superconductivity in order to make more conclusive statements about the sum rule behavior in the underdoped regime .
acknowledgments : we are very grateful to many people for the fruitful discussion and especially to n. bontemps , a. boris , c. bernhard and d. van der marel .
we would like to thank j. kortus for his help in the preparation of the paper .
this work was supported partially by rfbi under grants no.02 - 02 - 16658 and no.01 - 02 - 16719 , the russian program for the investigation of the superconductivity and the program of the presidium ras .
one of the authors ( e.g.m . ) is grateful to o.k .
anderson for the kind hospitality during the visit in the max - planck - institute fkf ( stuttgart ) where the part of this work has been done .
j.e .
hirsh , f. marsiglio , phys rev , * b62 * , 15131 ( 2000 ) m.r .
norman , m. randeria , b. janko , j .c .
campuzano , phys .
rev .
, * b61 * , 14742 ( 2000 ) d.j .
scalapino , s.r .
white , phys .
rev .
, * b58 , * 8222 ( 1998 ) a.j .
legget , phys .
rev .
lett . ,
* 83 * , 392 ( 1999 ) m.r .
norman , g. pepin , cond .-
mat./0201415 e.h .kim , phys .
rev .
, * b58 * , 2452 ( 1998 ) d.n .
basov _
et al . , _ science , * 283 * , 49 ( 1999 ) h.j.a .
molegraaf _ et al_. , science , * 295 * , 2239 ( 2002 ) m.j .
holcomb _
et al . , _ phys.rev . , * b53 * , 6734(1996 ) a.f .
santander - syro _ et al .
, _ cond .-
mat./0111539 n.n .
bogolyubov , v.v .
tolmachev , d.v .
shirkov , _ a new method in the theory of superconductivity _ , consultants bureau , n.y .
( 1959 ) j. bardeen , l.n .
cooper , j.r .
schrieffer , phys.rev . , * 108 * , 1175 ( 1957 ) m. tinkham , _ introduction to superconductivity _ ( mcgrow - hill , n.y.,1996 )
d.n .
basov _
et al . _
phys .
rev .
, * b63 * , 134514 ( 2001 ) n.w . ashcroft and m.d .
mermin , _ solid state physics _
( saunders , philadelphia , 1976 ) g.m .
eliashberg , zh .
exp .
theor .
phys . ,
* 38 * , 966 ; * 39 * , 1437(1960 ) ; sov .
phys .
jetp , * 11 * , 696 ( 1960 ) ; * 12 * , 1000 ( 1961 ) p. monthoux , a.v .
balatsky , d. pines , phys .
rev .
let .
* 67 * , 3448 ( 1991 ) s.b .
nam , phys.rev.,*156 * , 470 ( 1967 ) w. show , j.c .
swihart , phys .
rev .
lett . ,
* 20 * , 1000 ( 1968 ) s.v .
shulga , o.v .
dolgov , e.g. maksimov , physica * c 178 * , 266 ( 1991 ) e.g. maksimov , ufn * 170 * , 1033 ( 2000 ) [ physics - uspekhi , * 43 * , 965 ( 2000 ) ] e.g. maksimov _ et al . , _ sol .
st .comm . , * 112 * , 449 ( 1999 ) a. lanzara _ et al_. , nature , * 412 * , 510 ( 2001 ) h .- y .
kee _
et al . _
cond .-
mat./0110478 ar .
abanov _
et al . _
cond .-
mat./0112126 z .- x .
shen _ et al . _
cond .-
mat./0108381 s. schmitt - rink , c.m .
varma , a.e .
ruckenstein , phys .
rev .
lett . , * 63 * , 445 ( 1989 ) a.v .
puchkov _ et al_. , j. phys .
: cond . matter , * 8 * , 10049 ( 1996 ) |
although all experimentally observed particles are either bosons or fermions , the general principles of quantum mechanics do not prohibit existence of particles obeying other types of statistics @xcite .
the motivation to consider an exotic statistics is that it may provide an _ effective _ description of the dynamics of particles if their interaction is so strong that the available occupancy for a given state depends on the number of particles already present in the state @xcite .
an exotic statistics can also arise if particles possess a hidden internal degree of freedom that is invisible in the hamiltonian but that can become dynamically relevant .
finally , an exotic statistics can emerge in description of physical models in one or two spatial dimensions .
the reason is that in this case the multi - particle wave function is not necessarily even or odd and particles ( anyons ) may obey a fractional statistics @xcite .
moreover , in low - dimensional theories the exchange statistics of the fields present in a hamiltonian is not directly related to the exclusion statistics of the corresponding particles .
for instance , although particles in the one - dimensional real coupling affine toda field theories ( atft ) are bosons , one has to impose the pauli principle on their momenta in the thermodynamic bethe ansatz analysis in order to obtain correct values of corresponding central charges @xcite .
another example is the conformal field theory , where quasi - particle spectra can be constructed with use of an exotic exclusion statistics @xcite .
the first attempt to introduce an exotic statistics is attributed to g. gentile @xcite , who proposed an intermediate statistics , in which at most @xmath1 indistinguishable particles can occupy a given state .
this intermediate statistics interpolates between fermions and bosons , that are recovered for @xmath2 and @xmath3 . a decade after gentile s work h.s .
green introduced the parafermi statistics @xcite that also fixes maximal occupancy for a given state . since then , various aspects of systems governed by the gentile or parafermi statistics were discussed in the literature , among them possible applications to the elementary particles theory @xcite and to the quark confinement problem in the qcd @xcite .
many authors discussed thermodynamic properties of an ideal gas obeying the gentile or parafermi statistics @xcite . in the present paper
we will study the thermodynamic limit of one - dimensional relativistic integrable field models governed by a gentile - like statistics .
more precisely , the main purpose of this manuscript is to implement systematically a generalized extensive exclusion statistics into the thermodynamic bethe ansatz ( tba ) analysis of systems where the interaction of particles is relativistic , short - range and characterized by a factorizable scattering matrix .
the tba was originally developed in the seminal papers by yang and yang @xcite for treating a one - dimensional ideal gas . in
later works @xcite this technique was adopted to one - dimensional relativistic models .
the boundary condition for a many - particle wave function leads to what is commonly referred to as the bethe ansatz equation , that provides the quantization condition for possible momenta of the system .
the tba constitutes an interface between massive integrable models and conformal field theories . from the tba
one can extract information about the ultraviolet limit of the underlying massive model , in particular , find the corresponding ( effective ) central charge . in the derivation of the tba equation the underlying exclusion statistics is usually taken to be either of bosonic or fermionic type @xcite . in the present paper the tba equation will be derived and studied for significantly more general situation .
namely , let @xmath4 denote the total dimension of the hilbert space for a system of @xmath5 indistinguishable particles that can occupy @xmath6 different states in the fock - space .
assume that there exists a function such that its @xmath6-th power is a generating function for the dimensions @xmath4 , i.e. @xmath7 in the statistical mechanics , if the variable @xmath8 is understood as fugacity , the sum on the right hand side is the grand partition function of the system .
the property of a grand partition function to be an @xmath6-th power of a function independent of @xmath6 is called _ extensivity_. as we are going to demonstrate below , this property allows us to develop the tba analysis for a rather general choice of @xmath9 .
let us stress that for the purpose of deriving the tba equation even asymptotical extensivity suffices , i.e. , equation ( [ fgen ] ) should hold in the large @xmath6 limit . of course ,
if the explicit form of @xmath9 is known , we can use it to simplify the general formulae .
for instance , an instructive example is the gentile statistics of order @xmath1 which is defined by the following choice of @xmath9 : @xmath10 for @xmath2 and @xmath3 this statistics describes ordinary fermions and bosons , respectively .
although @xmath11 is the simplest non - trivial choice of @xmath9 , it has all features of a generalized extensive statistics
. therefore , the gentile statistics will be our main working example below .
the paper is organized as follows . in section 2
we describe possible types of a generalized extensive statistics and study properties of the corresponding entropy densities in the thermodynamic limit .
the gentile statistics and the @xmath0-ons statistics are discussed in this context . in section 3
we compare properties of the gentile statistics and the haldane - wu statistics . in section 4
we derive the thermodynamic bethe ansatz equation and the finite - size scaling function for a relativistic multi - particle system in which statistical interaction is governed by a generalized extensive statistics . in section 5
we derive the y - system related to such a generalized statistics . in section 6
we study the ultraviolet limit of the generalized tba equation and obtain an expression for the corresponding effective central charge . in section 7
we compute finite - size scaling functions and central charges related to some affine toda models for the gentile statistics and the @xmath0-ons statistics and compare them with their counterparts for the haldane - wu statistics .
our conclusions are stated in section 8 .
consider a system of @xmath5 indistinguishable particles that can occupy @xmath6 different states in the fock - space .
assume that the particles obey the gentile statistics of order @xmath1 , i.e. , that each state can be occupied by at most @xmath1 particles ( with @xmath1 being a positive integer number ) .
there are several combinatorial ways to compute the total dimension of the hilbert space for such a system .
for instance , we first choose @xmath12 states which are occupied by at least one particle , then among these @xmath12 states we choose @xmath13 states which are occupied by at least two particles , etc .
this way of counting yields the following expression for the total dimension of the hilbert space @xmath14 where @xmath15 are binomial coefficients ( and @xmath16 if @xmath17 or if @xmath18 ) .
now , let us compute @xmath6-th power of the polynomial @xmath11 defined in ( [ gf ] ) .
thanks to the obvious recursive relation @xmath19 , we can do this by @xmath1 consecutive applications of the binomial formula .
then it is easy to see that @xmath6-th power of @xmath11 is a generating function for the dimensions @xmath20 , i.e. @xmath21 thus , the gentile statistics is a particular case of an extensive statistics .
let us stress that the order @xmath1 of the gentile statistics must be a positive integer . although we could try to extend it to other values by formal replacement of ( [ gf ] ) with @xmath22 the resulting function @xmath11 will not satisfy the important positivity property ( see below ) .
consider now a more general system of @xmath5 indistinguishable particles that possesses , at least for large @xmath6 , the extensivity property ( [ fgen ] ) with @xmath23 where @xmath24 can be a finite positive number or infinity .
the taylor coefficients @xmath25 are fractional dimensions of levels in a single - state hilbert subspace and can be regarded as probabilities of occupation of the given state by @xmath26 particles ( see related discussions in @xcite ) .
the total dimension of the single - state subspace is @xmath27 .
it is natural to require _ positivity _ of the quantities
@xmath25 : @xmath28 the first condition here implies that the vacuum is realized with probability one independently of the size @xmath6 of the system .
furthermore , we assume that the series ( [ f ] ) converges in the complex plane for @xmath29 , where @xmath30 .
this implies that @xmath9 belongs to one of the following types : @xmath31 the types i , ii , and iii consist of finite degree polynomials , analytic functions with infinite taylor series , and meromorphic functions , respectively .
for example , a finite order gentile statistics belongs to the type i , the boltzmann statistics ( @xmath32 ) belongs to the type ii , and the bosonic statistics ( @xmath33 ) is of the type iiib .
the reason we divided the type iii into two subtypes will become clear below .
let us remark that , in addition to ( [ posit ] ) , one may require that @xmath34 with the motivation that the statistics should not be deformed if only one particle is present in the system . although all the concrete cases which we consider below do fulfill this requirement , it does not seem to be crucial for the general considerations .
the entropy of the system under consideration is @xmath35 , where @xmath26 is the boltzmann constant .
we need to compute the entropy in the thermodynamic limit , i.e. , when @xmath36 and @xmath37 is fixed .
the extensivity property ( [ fgen ] ) allows us to express @xmath4 as the following contour integral @xmath38 where @xmath39 , and we denoted @xmath40 .
now we can find an asymptotic expression for the entropy in the thermodynamic limit by application of the saddle point method to the integral ( [ cint ] ) .
the saddle point of the exponent in ( [ cint ] ) , denote it @xmath41 , is found as the positive root of the equation ( prime stands for a derivative ) @xmath42 choosing @xmath43 in ( [ cint ] ) , so that the saddle point belongs to the integration contour , we apply the standard result ( see , e.g. , @xcite ) for the laplace integral : @xmath44 where @xmath45 . for any @xmath9 satisfying ( [ posit ] ) and @xmath41 obeying ( [ xeq ] ) we have @xmath46 due to the hadamard theorem @xcite . from ( [ expr ] )
we obtain the entropy density @xmath47 for a given value of @xmath48 @xmath49 where @xmath41 is positive and satisfies ( [ xeq ] ) .
two remarks are in order here . strictly speaking , if @xmath9 is degenerated , i.e. , if there exists integer @xmath50 such that @xmath51 , then the above derivation modifies since @xmath9 has @xmath52 saddle points . however , the corresponding physical interpretation is simply that particles can be added only in @xmath52-tuples .
therefore , it is natural to introduce cluster variables @xmath53 , @xmath54 , etc . in terms of these new variables formulae ( [ cint])-([expr ] )
remain valid .
more important remark is that for deriving ( [ lim ] ) the property @xmath55 is crucial . in general , without the positivity property ( [ cint ] ) does not have a real and uniform in @xmath48 limit .
moreover , if we allow @xmath9 to have negative taylor coefficients , then some @xmath4 in the expansion of @xmath56 may also be negative .
this leads to an immediate problem with the definition of the entropy as a logarithm of @xmath4 .
however , if the first negative coefficient in the expansion of @xmath56 appears only at sufficiently large power @xmath57 ( of order @xmath58 ) , then @xmath9 violating the positivity condition can still have well defined entropy density for certain range of @xmath48 .
for instance , such is the case of the haldane - wu statistics with statistical interaction @xmath59 ( see section 3 ) , where @xmath60 .
however , in the present paper we will restrict our consideration , for simplicity , only to partition function strictly obeying the positivity condition .
the entropy density is a key quantity for developing the tba analysis .
from the definition of @xmath48 it is clear that @xmath61 for @xmath9 of the type i and @xmath62 for the other types .
formally this follows from equation ( [ xeq ] ) which shows that @xmath48 varies from @xmath63 to @xmath64 .
moreover , @xmath65 is a strictly increasing function because @xmath66 for @xmath67 . since @xmath9 is regular at zero , it is easy to see that @xmath68 tends to zero for small @xmath48 . together with the condition @xmath69 this yields @xmath70 . in order to analyze further behaviour of @xmath47
, we employ ( [ xeq ] ) and derive from ( [ lim ] ) that @xmath71 since @xmath72 , we conclude that @xmath73 , that is @xmath47 is a convex up function .
furthermore , ( [ der ] ) implies that @xmath74 .
note , however , that @xmath47 is not guaranteed to be positive for the whole range of @xmath48 .
more precisely , @xmath47 may have one root at the interval @xmath75 .
indeed , for @xmath9 of the type i we derive from ( [ xeq ] ) and ( [ lim ] ) that @xmath76 that is , @xmath47 is non - negative everywhere at @xmath61 only if the last taylor coefficient @xmath77 is greater or equal to one . for the other types of @xmath9 we have the following properties @xmath78 indeed , in the first case @xmath79 , hence , according to ( [ der ] )
, @xmath47 is strictly increasing . in the second case
we have @xmath80 .
the integral here apparently diverges since it exceeds @xmath81 , where we can choose any @xmath82 such that @xmath83 .
investigation of thermodynamic properties of models in which the entropy density becomes negative is rather problematic .
therefore eqs .
( [ sd])-([sinf2 ] ) suggest that we should restrict the choice of @xmath9 to the type i with @xmath84 ( the gentile - like statistics ) and to the type iiib ( the bose - like statistics ) . however , let us notice that in the tba framework the variable @xmath41 depends on the rapidity @xmath85 .
for some relativistic models with factorizable scattering the typical picture @xcite is that @xmath86 falls off as @xmath87 grows ( for instance , @xmath88 for the trivial s - matrix ) . in this case
@xmath89 and we can consider @xmath9 of the types ii and iiia if @xmath47 remains non - negative for @xmath90 . in the context of this discussion
it is instructive to consider the so - called @xmath0-ons @xcite .
these are particles with statistics interpolating between fermions ( @xmath91 ) and bosons ( @xmath92 ) in the following simple way @xmath93 solving equation ( [ xeq ] ) for this choice of @xmath48 it is easy to find the corresponding single - state partition function : @xmath94_\gamma \frac{t^k}{k ! } \,,\ ] ] where @xmath95_\gamma \= \prod_{m=1}^{k } ( 1 \- m\gamma)$ ] . for @xmath96 the positivity condition ( [ posit ] )
is fulfilled only if @xmath97 , with @xmath24 positive integer . in this case
@xmath98 is of the type i , the corresponding maximal occupancy is @xmath99 . since @xmath100 for @xmath101 , the entropy density becomes negative for certain range of @xmath48 .
the case of @xmath102 describes the boltzmann statistics . here
@xmath103 belongs to the type ii and @xmath47 is negative for @xmath104 .
finally , for @xmath105 the positivity condition ( [ posit ] ) is always fulfilled and @xmath98 is of the type iiia or iiib depending on the value of @xmath0 .
returning to the gentile statistics , we can summarize the properties of the corresponding entropy density @xmath106 as follows ( see fig . 1 for illustration ) .
@xmath106 is a convex up function defined for @xmath107 and vanishing at the end points of this interval .
it attains the maximum at @xmath108 : @xmath109 furthermore , @xmath106 possesses the following symmetry @xmath110 this is a consequence of the equality @xmath111 which , in turn , follows from the relation ( [ gen ] ) .
it is interesting to compare thermodynamic properties of a system of particles obeying the gentile statistics with those of a system obeying the so - called haldane - wu statistics . for the latter ,
the total dimension of the hilbert space is postulated to be @xcite @xmath112 this expression interpolates between the bosonic ( @xmath113 ) and the fermionic ( @xmath114 ) statistics .
introduction of such an interpolating statistics was motivated by haldane s generalization @xcite of the pauli exclusion principle to the form @xmath115 here @xmath116 is the number of available states ( holes ) before the @xmath5-th particle has been added to the system .
the parameter @xmath59 is called the statistical interaction .
properties of an ideal gas obeying the haldane - wu statistics were actively discussed in the literature @xcite . the corresponding tba equation for relativistic integrable models was obtained in @xcite and investigated in @xcite .
a direct comparison of the quantities @xmath20 and @xmath117 is problematic if the system contains a finite number of particles . indeed ,
formula ( [ gcount ] ) has a bulky form for generic @xmath1 ( i.e. , for @xmath118 ) .
moreover , it requires additional conventions to make exact sense of expression ( [ wahr ] ) for generic @xmath59 ( i.e. , for @xmath119 ) .
actually , there exists no prescription for counting of states on the microscopical level that would lead to ( [ wahr ] ) ( see related discussions in @xcite ) . to overcome these difficulties we can compare the two statistics in the thermodynamic limit .
more precisely , let us compare the entropy density @xmath106 with its counterpart @xmath120 that is readily derived from ( [ wahr ] ) with the help of the stirling formula @xmath121 let us notice that the haldane - wu statistics is extensive but the corresponding function @xmath122 does not satisfy the positivity property ( [ posit ] ) .
indeed , we can reconstruct @xmath122 first as a function of @xmath48 by the following formula @xmath123 that follows from ( [ lim ] ) and ( [ der ] ) . for @xmath124 given by ( [ limhw ] )
this yields @xmath125 which is a well - defined function . but
reexpressing it in terms of @xmath8 with the help of ( [ xeq ] ) we will always obtain a function that breaks the positivity property ( [ posit ] ) .
for instance , for @xmath126 we find ( which coincides with the result of @xcite ) .
however , the first negative term in expansion of the @xmath6-th power of this series appears for @xmath127 .
this implies that in the large @xmath6 limit thermodynamic quantities are well - defined for @xmath128 . for our purposes we can regard ( [ limhw ] ) as _ a definition _ of the haldane - wu statistics .
then we have a finite maximal occupancy @xmath129 for a given number of particles @xmath6 simply because @xmath124 is defined for @xmath130 .
furthermore , @xmath124 is a convex up function vanishing at the end points of this interval .
thus , we see that properties of the entropy density @xmath124 are similar to those of a type i extensive statistics and , specifically , to those of the gentile statistics .
therefore , it is natural to ask how much the behaviour of @xmath106 differs from @xmath124 if @xmath131 ( when the corresponding maximal occupancies coincide ) . first ,
notice that the function @xmath132 can not coincide with @xmath106 identically for generic @xmath1 since @xmath124 does not possess a symmetry like ( [ sym ] ) if @xmath133 .
next , we can compare the mid - point value @xmath134 with the corresponding value of @xmath106 given by ( [ gmax ] ) .
taking into account the inequalities @xmath135 t \ln t < ( t-1)\ln(t+1 ) \qquad { \rm for } \quad 0<t<1 \ , , \end{array}\ ] ] we establish that @xmath136 for @xmath137 .
actually , numerical computations for various values of @xmath137 show that @xmath106 majorizes @xmath132 _ everywhere _ at @xmath138 , @xmath139 for illustration , the case of @xmath140 is presented in fig . 1 .
thus , we see that for generic @xmath1 the difference between @xmath132 and @xmath106 is not small if @xmath48 is not close to @xmath141 or @xmath142 .
it is instructive to compare these functions also near the end points ( in particular , this will provide an additional support to the assertion ( [ ineq2 ] ) ) . in a vicinity of @xmath141 equation ( [ xeq ] ) is solved as @xmath143 , and we find that @xmath144 , where @xmath145 stands for the kronecker symbol . comparing this expansion with @xmath146 which follows from ( [ limhw ] )
, we infer that for small @xmath48 functions @xmath124 and @xmath106 take close values ( up to the order @xmath147 ) for _ any choice _ of @xmath1 and @xmath59 . actually , employing ( [ xeq ] ) , it is easy to derive the following expansion for the entropy density of an arbitrary extensive statistics with @xmath9 satisfying ( [ posit ] ) : @xmath148 thus , for small @xmath48 in the thermodynamic limit , the haldane - wu and gentile statistics are close not only to each other but to any extensive statistics for which two first taylor coefficients of @xmath9 are @xmath149 .
for instance , for small @xmath48 the haldane - wu and gentile statistics are close also to the statistics of @xmath0-ons ( [ fga ] ) .
consider now @xmath150 for small positive @xmath151 .
then , employing ( [ exp ] ) and the symmetry ( [ sym ] ) , we derive : @xmath152 . on the other hand , setting @xmath153 in ( [ limhw ] ) , we obtain @xmath154 .
we see that in a vicinity of @xmath142 the difference between @xmath106 and @xmath132 is not even of first order in @xmath151 but of order @xmath155 ( so that the corresponding plots are visibly different here , see fig . 1 ) .
summarizing , it appears plausible that the gentile statistics of order @xmath156 _ majorizes _ the haldane - wu statistics with parameter @xmath131 and these statistics are _ not close _ except for small values of @xmath157 .
now we will consider a relativistic multi - particle system of @xmath158 different species of particles confined to a finite region of the size @xmath159 .
we denote by @xmath160 the number of particles and by @xmath161 the dimension of the fock - space related to the species @xmath162 . in the thermodynamic limit both the dimensions of the fock - space and the size @xmath159 of the region were the system is confined approach infinity but the ratios @xmath163 remain finite @xcite . for the fraction of particles with rapidities between @xmath164 and @xmath165
it is convenient to introduce densities @xmath166 as usual , the rapidity @xmath167 parameterizes the two - momentum @xmath168 and the total energy of the system is given by @xmath169 = l \sum_{a=1}^l \int_{-\infty}^{\infty } { \rm d}\theta \ , \rho_a^r ( \theta ) \ , m_a \cosh \theta \,.\ ] ] according to ( [ dens ] ) the variable @xmath48 in ( [ lim ] ) is now related to the particle densities @xmath170 thus , we can express the entropy as the following functional @xmath171 = kl \sum_{a=1}^l \int_{-\infty } ^{\infty } { \rm d}\theta \ , \rho_a(\theta ) s(\mu_a ) = kl \sum_{a=1}^l \int_{-\infty } ^{\infty } { \rm d}\theta \ , \bigl\ { \rho_a(\theta ) \ln f_a(x_a(\theta ) ) - \rho_a^r(\theta ) \ln x_a(\theta ) \bigr\ } \,,\ ] ] where each @xmath172 satisfies ( [ xeq ] ) and hence it is a function of @xmath173 .
the subscript of @xmath174 means that different species may have different single - state partition functions . in the formulation of the tba equation for ordinary statistics
it is common to introduce the so - called pseudo - energies @xmath175 such that bosonic ( upper sign ) and fermionic ( lower sign ) distributions are @xmath176 . in view of ( [ xeq ] ) , it is natural to define the pseudo - energies for an arbitrary extensive statistics as follows @xmath177 so that @xmath178 .
for instance , the distribution for the gentile statistics acquires the following form in terms of the pseudo - energies ( with a slightly different meaning it was introduced by gentile @xcite for an ideal gas ) @xmath179 according to the fundamental principles of thermodynamics the equilibrium state of a system is found by minimizing the free energy @xmath180 .
hence , keeping the temperature constant we obtain the equilibrium condition by minimizing @xmath181 = e\left [ \rho^r \,\right ] -ts\left [ \rho , \rho^r \,\right ] $ ] with respect to @xmath182
. the equilibrium condition reads @xmath183 the admissible momenta in the system are restricted by the boundary conditions . to derive the corresponding quantization equations one takes a particle in the multi - particle wave function on a trip through the whole system @xcite .
the particle will scatter with all other particles , which yields the following set of equations determining admissible rapidities @xmath184 where @xmath185 , and @xmath186 is a two - particle scattering matrix .
the constant phases @xmath187 on the r.h.s . of ( [ ba ] ) may depend on the particle s species ; for our purposes their exact values are irrelevant .
taking the logarithmic derivative of ( [ ba ] ) and employing densities as in ( [ dens ] ) , we obtain @xmath188 here @xmath189 stands for the convolution and we introduced as usual the notation @xmath190 . employing equations ( [ xeq ] ) , ( [ lim ] ) and ( [ mu ] ) , we compute variations of the entropy ( [ entro ] ) : @xmath191 substitution of these relations into the equilibrium condition ( [ equ ] ) together with ( [ bou ] ) yields
the desired thermodynamic bethe ansatz equations for a system in which statistical interaction between particles of @xmath162-th species is described by a single - state partition function @xmath174 , @xmath192 in general , this set of integral equations can not be solved analytically .
however , one can try to solve these tba equations numerically with the help of the iteration method ( as we will do for some examples in section 7 ) . for the ordinary statistics and the haldane - wu case it is known that such an iterative procedure converges provided that @xmath193 falls off sufficiently fast .
the most complete proof of this assertion is presented in @xcite and is based on the fixed point theorem .
presumably this proof extends to the case of ( [ gtba ] ) ( although some further restrictions on @xmath194 may be required ) , but we will not discuss this here . let us notice only that if ( [ gtba ] ) can be solved iteratively and @xmath195 is symmetric in @xmath85 ( which is always the case for a unitary s - matrix ) , then @xmath172 is also symmetric in @xmath85 .
we now substitute the expressions for the total energy ( [ ener ] ) and the entropy ( [ entro ] ) together with equations ( [ bou ] ) and ( [ gtba ] ) back into the expression for the free energy , and obtain @xmath196 the relation between the free energy and the finite - size scaling function is well - known to be @xmath197 @xcite .
as usual we introduce a variable @xmath198 and set now the boltzmann constant to be one .
then the finite - size scaling function acquires the form @xmath199 here we assumed that @xmath200 .
once more for @xmath201 and @xmath202 we recover the well - known expressions for the fermionic and bosonic scaling functions .
it is instructive to compare the above tba equation ( [ gtba ] ) and the finite - size scaling function ( [ c(r ) ] ) with those for the haldane - wu statistics @xcite : @xmath203 @xmath204 here @xmath205 , and @xmath59 is a matrix which appears on the r.h.s . of ( [ pauli ] ) if we consider a system of several species .
as was discussed in @xcite , the tba equation ( [ hwtba ] ) leads to an equivalence principle : two systems having the same mass spectra and identical quantities @xmath206 are thermodynamically equivalent ( as seen from ( [ hwcr ] ) , their finite - size scaling functions coincide ) .
our tba equation ( [ gtba ] ) does not possess such a feature .
more precisely , for ( [ gtba ] ) there exists no way to compensate a difference in statistics by a change of @xmath195 independent of the explicit form of the s - matrix .
for some class of models it is possible to carry out certain manipulations on the tba equations such that the original integral tba equations acquire the form of a set of functional equations in new variables @xmath207 @xcite . these functional equations (
commonly referred to as ) have the further virtue that unlike the original tba equations they _ do not _ involve the mass spectrum . in the @xmath208-variables certain periodicities in the rapidities are exhibited more clearly .
these periodicities may then be utilized in order to express the quantity @xmath209 as a fourier series , which in turn is useful to find solution of the tba equations and expand the scaling function as a power series in the scaling parameter @xmath210 .
we will now demonstrate that similar equations may be derived for a multi - particle system in which the dynamical scattering is governed by the scattering matrix related to the ade - affine toda field theories and the statistical interaction is of a general type considered above .
consider the minimal part ( i.e. , independent of the coupling constant ) of the scattering matrix of the ade - affine toda field theories . as was shown in @xcite , upon appropriate choice of cdd - ambiguities ,
these s - matrices satisfy the identity @xmath211 where @xmath212 denotes the coxeter number , @xmath158 the rank and @xmath213 the incidence matrix of the corresponding lie algebra .
it is then straightforward to derive the `` y - system '' @xmath214 where @xmath215 and @xmath216 equations ( [ ys ] ) follow upon first adding ( [ gtba ] ) at @xmath217 and @xmath218 and subtracting @xmath213 times ( [ gtba ] ) at @xmath167 from the sum .
thereafter we employ the fact @xcite that the masses of an ade affine toda field theory are proportional to the perron - frobenius vector of the corresponding cartan matrix , i.e. , @xmath219 .
then , with the help of ( [ rad ] ) , equations ( [ ys ] ) follow . in comparison with ( [ gtba ] ) equations ( [ ys ] ) have already the virtue that they are simple functional equations and do not involve the mass spectrum . for the gentile statistics we have @xmath220 . in particular , for @xmath221 we have @xmath222 and recover the known fermionic y - system .
the bosonic case ( @xmath3 ) corresponds to @xmath223 . and the case of @xmath224 seems to be particularly interesting since here @xmath225 possesses an additional symmetry @xmath226 or , equivalently , @xmath227 .
in the small @xmath210 limit the scaling function @xmath228 becomes the effective central charge of a conformal field theory @xcite which describes the ultraviolet limit of a given massive model .
that is , @xmath229 , where @xmath230 is the conformal anomaly and @xmath231 is the lowest conformal weight in the corresponding conformal field theory . in order to evaluate @xmath228 in the @xmath232 limit
we substitute the derivative of ( [ gtba ] ) into ( [ c(r ) ] ) , replacing in both equations @xmath233 by @xmath234
. then we integrate the resulting equation by parts , assuming that @xmath195 is symmetric and falls off sufficiently fast ( typically , @xmath235 as @xmath236 ) .
next we substitute back equation ( [ gtba ] ) .
finally we change variables from @xmath85 to @xmath86 , taking into account that @xmath237 ( as follows from ( [ gtba ] ) ) , and again integrate by parts .
all this yields @xmath238 @xmath239 here we denoted @xmath240 .
these quantities can be determined from the _ constant _ tba equations that follow from ( [ gtba ] ) when @xmath241 ( this derivation requires that @xmath172 become constant in a region of @xmath85 of order @xmath242 when @xmath210 is small ; see more detailed explanations in @xcite ) @xmath243 where @xmath244 . as a simple example we consider the statistics of @xmath0-ons ( [ fga ] ) , assuming for simplicity that @xmath0 is the same for all species ( generalization to different @xmath245 is obvious ) .
here the integral in ( [ c2 ] ) can be computed explicitly in terms of the euler dilogarithm which is defined for @xmath246 as @xmath247 .
we have @xmath248 and hence @xmath249 where @xmath250 satisfies ( [ ctba ] ) .
since @xmath251 , we can rewrite ( [ cga1 ] ) for positive @xmath0 employing the rogers dilogarithm which is defined as @xmath252 for @xmath253 .
then , using the abel identity @xmath254 , we obtain @xmath255 notice that this expression does not diverge for small @xmath0 ; @xmath256 is the value corresponding to the boltzmann statistics .
consider now the gentile statistics , assuming that the order @xmath1 is the same for all species . in this case
we also can compute the integral in ( [ c2 ] ) explicitly : @xmath257 and we see that for the gentile statistics expression ( [ c2 ] ) acquires the following nice form involving only the rogers dilogarithms : @xmath258 introducing , in agreement with ( [ id1 ] ) , the constants @xmath259 , we obtain from ( [ cg ] ) for @xmath260 and @xmath221 @xmath261 these are the well - known values for the bosonic and fermionic statistics @xcite . notice that in the latter case one usually derives @xmath262 .
this is equivalent to our formula ( [ cbcf ] ) due to the abel identity .
in fact , the abel identity can be used in a similar way for any @xmath1 of the form @xmath263 . in this case
we can rewrite ( [ cg ] ) as follows @xmath264 in this form @xmath265 is a manifestly positive and monotonic function of @xmath250 ( the dilogarithm @xmath266 grows monotonically for @xmath253 ) .
actually , this property holds not only for the gentile statistics but in general case as well . to prove this statement we take a derivative of ( [ c2 ] ) and use ( [ xeq ] ) and the formula ( [ lim ] ) for the entropy density .
this yields @xmath267 as we discussed in section 2 , we should consider only such models where @xmath268 . in this case the r.h.s
. of ( [ dc ] ) is positive .
hence , @xmath265 is always a monotonically increasing function of @xmath250 ( and therefore a monotonically decreasing function of @xmath269 ) .
moreover , taking into account ( [ posit ] ) , we infer from ( [ c2 ] ) that @xmath270 , thus we also proved positivity of @xmath271 .
since these properties are crucial for a physical interpretation of @xmath265 , let us underline that they hold only if the first condition in ( [ posit ] ) is fulfilled .
let us notice also that the properties of @xmath265 together with ( [ ctba ] ) imply that for any model such that @xmath272 ( which is the case , for instance , for all the ade - affine toda models @xcite ) the value of @xmath265 does not exceed that of the corresponding free model ( where all @xmath273 and consequently all @xmath274 ) .
hence , ( [ c2 ] ) evaluated for @xmath274 yields the upper bound on @xmath265 , @xmath275 notice , however , that for a statistics of the type iiib with @xmath276 we have @xmath277 , so that in this case the upper bound on @xmath265 will be lower than ( [ cmax ] ) . from a physical point of view
this implies that such a statistics can not emerge in models with weak interaction and , in particular , in a free model .
it was argued in @xcite that the quantity @xmath278 ( for the ordinary statistics ) can be interpreted as a massless degree of freedom associated to the @xmath162-th species in the massive model .
the properties of @xmath265 which we proved above show that the interpretation of @xmath278 can be extended to the case of a generalized extensive statistics as well . in this context , it is important to remark that for the ordinary statistics ( actually , for the gentile statistics of any order ) we have @xmath279 as follows from ( [ cg ] ) ( recall that @xmath280 ) if all @xmath281 are non - negative . however , for a general statistics the upper bound ( [ cmax ] ) on @xmath278 depends on the choice of @xmath174 and may not equal to one .
for instance , taking @xmath282 with @xmath283 , we obtain @xmath284 for the corresponding free theory .
in such a case , the massless degree of freedom of the corresponding free particle exceeds that of a free boson .
this can possibly imply that we have to restrict the choice of statistics to such @xmath174 that @xmath285 .
the most simple examples which illustrate the features outlined above more concretely are the ising model ( @xmath286-minimal affine toda field theory ) with @xmath287 and the klein - gordon model with @xmath288 in both cases equation ( [ gtba ] ) is solved trivially , @xmath289 .
then for the gentile statistics ( [ gf ] ) we can compute the entire scaling function ( [ c(r ) ] ) : @xmath290 where @xmath291 is the modified bessel function .
we depict @xmath292 in fig . 2 , referring to them by a slight abuse of notation as @xmath293 .
one observes that the difference in statistics affects most severely the ultraviolet region .
the scaling functions for different @xmath1 converge relatively fast towards each other in the infrared regime .
these appear to be common features of a general deformed statistics . as seen from fig . 2 ,
the behaviour of @xmath292 in the ultraviolet region ( there is no first order term in the small @xmath210 expansion of @xmath292 ) is similar for any finite @xmath1 to that of the fermionic scaling function and is different from the bosonic case ( where the first order term is present ) . to explain this we observe that , as follows from ( [ cf ] ) , @xmath294 where @xmath295 is the bosonic scaling function .
now it is obvious that the first order terms on the r.h.s . of ( [ cb ] )
cancel each other for any finite @xmath1 .
the well - known property of the asymptotic behaviour of the modified bessel function , @xmath296 , allows us to derive from ( [ cf ] ) @xmath297 which is in agreement with ( [ cg ] ) since @xmath298 now .
it is worth noticing that @xmath299 can be identified for every positive integer @xmath1 as the effective central charge of a minimal conformal model @xmath300 ( such that @xmath301 ) . in particular , @xmath302 correspond to the @xmath303 , @xmath304 , @xmath305 , @xmath306 minimal models , respectively . for the @xmath0-ons ( [ fga ] ) the scaling function related to the @xmath293-models
can also be computed : @xmath307 using again the asymptotic behaviour of the modified bessel function , we obtain @xmath308 in agreement with ( [ cga1 ] ) . here , unlike in the gentile case , it is not easy to see whether the effective central charge @xmath309 has rational value for a given @xmath0 . as we discussed in section 2 , positive @xmath0 is to be of the form @xmath310 with @xmath24 positive integer .
it appears then that the only @xmath24 leading to rational value of @xmath309 is @xmath311 . on the other hand
, @xmath0 varies continuously in the range @xmath312 , so that any rational value of @xmath309 not exceeding @xmath313 occurs here . for the boltzmann statistics @xmath314
is irrational .
finally , @xmath315 is the case of a type iiib statistics with @xmath276 , which , as we discussed above , can not emerge in a free model .
next we consider the scaling potts model , which was studied previously in the tba framework for the fermionic statistics in @xcite and for the haldane - wu statistics in @xcite .
the two particles in the model are conjugate to each other and consequently their masses are the same , @xmath316 .
the conjugate particle occurs as a bound state when two particles of the same species scatter .
the s - matrix of the scaling potts model equals the minimal s - matrix of the @xmath317-affine toda field theory .
the corresponding quantities @xmath195 that enter the tba equation are given by @xmath318 consider the scaling potts model with both species of particles obeying the gentile statistics of the same order @xmath1 . in this case the @xmath319-symmetry of the model is preserved so that @xmath320 and ( [ gtba ] ) reduces to a single integral equation .
it appears to be impossible to find an analytic solution @xmath86 to this equation , but it is straightforward to solve it numerically . taking @xmath321}(\theta )
= \exp\{-rm\cosh \theta\ } $ ] as the first approximation , we can iterate ( [ gtba ] ) as follows @xmath322}(\theta ) \right ) = -rm \cosh \theta - \frac{2\sqrt{3}}{\pi } \int_{-\infty } ^{\infty } { \rm d}\theta^\prime \frac{\cosh ( \theta -\theta^\prime ) } { 1 + 2 \cosh 2(\theta -\theta^\prime ) } \ln \left ( f_{\scriptscriptstyle g } \bigl ( x^{[n]}(\theta^\prime ) \bigr ) \right).\ ] ] convergence is achieved relatively quickly ( it depends on the value of @xmath323 ) .
the results for @xmath2 ( which is in complete agreement with the calculation in @xcite ) and for @xmath224 are shown in fig .
3 . to make contact with the literature
, we introduced the quantity @xmath324 .
one observes the typical plateau @xmath325 for some region of @xmath85 when @xmath326 , which is required to derive ( [ ctba ] ) .
we see also that for the same value of @xmath323 the functions @xmath327 corresponding to different @xmath1 have similar profiles but different heights of the plateau .
having solved the generalized tba equation , we can substitute @xmath86 into ( [ c(r ) ] ) and compute the entire scaling function @xmath328 .
the result of the numerical computation is shown in fig .
the behaviour of the scaling functions is similar to what we have already observed in the @xmath293 case . due to the fact that the scattering matrices of the scaling potts model and the scaling yang - lee model are related as @xmath329 , the generalized tba equations for the two models coincide .
the only difference is that the scaling yang - lee model has only one particle .
therefore its scaling function equals a half of that for the scaling potts model .
in section 3 we have seen that the entropy density corresponding to the gentile statistics majorizes that for the haldane - wu statistics .
it appears ( at least for certain models ) that an analogous relation holds for the corresponding finite - size scaling functions .
specifically , if @xmath292 is the scaling function ( [ c(r ) ] ) of @xmath293 or @xmath330 minimal affine toda model where all species obey order @xmath1 gentile statistics and @xmath331 is the scaling function ( [ hwcr ] ) of the same model where the haldane - wu statistical interaction is of the form @xmath332 , then for all @xmath333 we have @xmath334 for the @xmath293 case we can rigorously prove relation ( [ crin ] ) since the solutions of the generalized tba equations ( [ gtba ] ) and ( [ hwtba ] ) are found explicitly : @xmath335 this equation allows us to express @xmath86 in terms of @xmath336 : @xmath337 moreover , since @xmath338 , we infer from ( [ xy0 ] ) that @xmath339 as seen from ( [ c(r ) ] ) and ( [ hwcr ] ) , in order to establish ( [ crin ] ) it suffices to show that @xmath340 .
the latter relation is equivalent to the inequality @xmath341 as can be verified by simple algebraic manipulations using formula ( [ xy1 ] ) .
now it is elementary to check that ( [ yy ] ) is valid provided that ( [ yy0 ] ) holds .
thus , we proved the assertion ( [ crin ] ) in the @xmath293 case . for illustration ,
the scaling function @xmath331 for @xmath342 haldane - wu statistics is shown in fig .
2 . it lies everywhere below @xmath292 for @xmath224 gentile statistics with the maximal difference reached in the ultraviolet region . for a higher rank case , already for @xmath330 ,
the generalized tba equations are integral and it is not clear whether we can compare @xmath172 and @xmath343 analytically .
however , numerical computations for various @xmath1 show that ( [ crin ] ) holds in the @xmath330 case as well . for illustration , we present in fig .
4 the scaling function @xmath331 for the haldane - wu statistics with @xmath344 .
a particular consequence of ( [ crin ] ) is the following inequality for the corresponding central charges : @xmath345 here @xmath299 is found from ( [ c0 ] ) , ( [ ctba ] ) and ( [ cg ] ) whereas @xmath346 is the effective central charge related to the haldane - wu statistics with @xmath332 and is given by @xcite @xmath347 again , in the @xmath293 case we can prove relation ( [ cin ] ) directly . here
@xmath299 has a simple form ( [ cgt ] ) and therefore ( [ cin ] ) reduces to @xmath348 where @xmath349 is defined as in ( [ yy0 ] ) .
notice that @xmath350 for @xmath133 ( since @xmath351 would contradict ( [ gin ] ) ) .
therefore , @xmath352 .
using now a property of the dilogarithm , @xmath353 we establish the desired inequality ( [ din1 ] ) .
as we discussed above , the order @xmath1 of the gentile statistics is to be a positive integer .
however , if we regard ( [ c0 ] ) , ( [ ctba ] ) and ( [ cg ] ) as a _ definition _ of the quantity @xmath299 , then ( [ din1 ] ) extends to any @xmath354 ( since the proof does not require that @xmath1 be an integer ) as shown in fig .
5 . furthermore ,
considering @xmath1 in the range @xmath355 , we can prove with the help of ( [ gin ] ) that here @xmath356 and hence @xmath357 due to ( [ lin ] ) .
thus , for this range of @xmath1 we have a _
reverse _ inequality , i.e. , @xmath358 as we will see below , this inequality holds for higher rank cases as well . but let us stress again that @xmath299 in ( [ cin2 ] ) is a formally defined quantity which is not related directly to the gentile statistics . the inequality for the central charges ( [ cin ] ) is a weaker statement than the inequality for the scaling functions ( [ crin ] ) .
( [ cin ] ) involves dilogarithms , which makes its proof more complicated ( already in the @xmath293 case we used a rather subtle property ( [ lin ] ) ) .
but the advantage of ( [ cin ] ) for a numerical verification is that it relates numbers and not functions as ( [ crin ] ) .
moreover , since @xmath228 is a continuous function , validity of ( [ cin ] ) implies that ( [ crin ] ) holds at least in some ultraviolet region .
therefore , we will discuss below validity of ( [ cin ] ) for several affine toda models . as a mathematical by - product
, this provides us with interesting dilogarithm inequalities . in the @xmath330 case , eq .
( [ cin ] ) contains dilogarithms in a more involved way . as seen from ( [ ctba ] ) , ( [ cg ] ) and ( [ chw1 ] ) , it is equivalent to the following inequality @xmath359 here @xmath360 and @xmath361
are determined from the following equations that follow from ( [ ctba ] ) and ( [ chw1 ] ) upon noticing that @xmath362 @xmath363 it is worth remarking that , as was noticed in @xcite , @xmath364 coincides with the effective central charge of the calogero - sutherland model with the coupling constant @xmath365 .
numerical computations show that ( [ din2 ] ) indeed holds for any @xmath119 , that is for any @xmath366 ( here again we need not to restrict @xmath1 to be an integer ) .
furthermore , like in the @xmath293 case , inequality ( [ din2 ] ) reverses for @xmath355 , i.e. , for @xmath367 . for illustration ,
we depict the corresponding @xmath299 and @xmath346 in fig . 5 .
this result provides an additional support to our claim that in the @xmath330 case the entire scaling function @xmath292 majorizes @xmath331 for @xmath137 . it is interesting to remark
that , as follows from ( [ cg ] ) and ( [ chw1 ] ) , the bosonic version ( @xmath3 or , equivalently , @xmath368 ) of the @xmath330-minimal affine toda model has @xmath369 .
that is , then each species in the scaling potts model ( or the only species in the scaling yang - lee model ) appears to be a free fermion , whereas the entire central charge is that of a free boson ( see also related comments in @xcite ) .
that is why all curves on fig .
5 converge to the same value @xmath370 .
now we discuss briefly the simplest minimal affine toda models which have particle species with different masses .
we will compare numerical values of @xmath299 and @xmath371 again setting @xmath372 .
for this purpose we use ( [ ctba ] ) and ( [ chw1 ] ) , taking into account @xcite that the @xmath6-matrix in these equations is given by @xmath373 , where @xmath374 stands for the incidence matrix of the corresponding lie algebra * g*. in the @xmath375 case the masses of the three species are @xmath376 .
then , computing @xmath299 by formulae ( [ ctba ] ) and ( [ cg ] ) , we find : @xmath377 , @xmath378 , @xmath379 , @xmath380 for @xmath381 , respectively .
the values of @xmath371 corresponding to the haldane - wu statistics with @xmath332 were found in @xcite : @xmath382 , @xmath378 , @xmath383 , @xmath380 for @xmath384 , respectively .
carring out analogous computations for the @xmath385 case , where the masses of the four species are @xmath386 , @xmath387 , we find ( gentile statistics ) : @xmath388 , @xmath389 , @xmath390 , @xmath391 for @xmath381 , respectively ; and ( haldane - wu statistics ) : @xmath392 , @xmath389 , @xmath393 , @xmath391 for @xmath384 , respectively .
one can also consider a folded algebra case with two different masses , namely @xmath394 , where @xmath395 .
in general , one assigns to the @xmath396 minimal affine toda model a tad - pole graph @xcite such that the corresponding incidence matrix differs from that of @xmath397 only by 1 in the lower - right entry .
then one can check that @xmath250 in ( [ ctba ] ) and @xmath398 in ( [ chw1 ] ) coincide with those for the @xmath399 case ( a well - known fact for the fermionic statistics @xcite ) .
consequently , for any statistics we have @xmath400 .
therefore , the data for @xmath394 follow from the above results for @xmath385 . finally ,
in the @xmath401 case , where the masses of the four species are @xmath402 , we obtain for the gentile statistics : @xmath403 , @xmath378 , @xmath404 , @xmath405 for @xmath381 , respectively ; and for the haldane - wu statistics : @xmath406 , @xmath378 , @xmath407 , @xmath405 for @xmath384 , respectively .
we see that all the above results are in agreement with ( [ cin ] ) and ( [ cin2 ] ) .
this , together with the more detailed results we obtained in the @xmath293 and @xmath330 cases , allows us to conjecture that these inequalities and the inequality ( [ crin ] ) for the scaling functions hold actually for all simply - laced minimal affine toda models .
we remark also that in all the computations of @xmath408 in this subsection we always have the rule that the value of @xmath265 is smaller for the species which has heavier particle ( this was known for the fermionic statistics @xcite ) .
this observation provides an additional support to the above discussed interpretation of @xmath409 as a massless degree of freedom of the corresponding particle .
summarizing , we considered possible types of a generalized extensive statistics and established properties of the corresponding entropy densities . in particular , we established numerically ( and gave some supporting analytical arguments ) that the entropy density for the gentile statistics of order @xmath137 majorizes that for the haldane - wu statistics if they correspond to the same maximal occupancy .
further , we derived the thermodynamic bethe ansatz equation and the finite - size scaling function @xmath228 for a relativistic multi - particle system obeying a generalized extensive exclusion statistics .
we put particular emphasis on the ultraviolet limit of such a system .
we derived an expression for an effective central charge in the general case and showed that for the gentile statistics it acquires an elegant form involving dilogarithms .
we discussed a physical interpretation of the ` partial ' central charges @xmath265 and argued that it possibly leads to restrictions on the choice of statistics .
finally , we observed ( and partially proved ) that the majorizing properties of the gentile statistics with respect to the haldane - wu statistics extend also to finite - size scaling functions and central charges related to the minimal simply - laced affine toda models . in the presented analysis of thermodynamic properties of a relativistic multi - particle system the single - state partition function @xmath9 played a key role .
for a given model , if we know @xmath9 ( at least in the thermodynamic limit ) and the quantities @xmath195 , then we can compute , at least in principle , the entire scaling function which provides complete thermodynamic description of a system . in principle
, @xmath9 should be determinable from the corresponding hamiltonian .
however , this might be a non - trivial task since the exclusion statistics of particles can differ from the exchange statistics of the fields present in the hamiltonian .
therefore , it may be more practical for a given system to fix an exotic statistics a - priory and investigate whether it leads to realistic physical properties .
in particular , one can obtain information about the ultraviolet limit and see if it corresponds to an appropriate conformal field theory . as we have seen above , the different types ( [ types ] ) of @xmath9 lead to significantly different thermodynamic properties .
the question , which of these statistics can emerge in physical models , remains open .
another open problem is to prove rigorously the assertion ( [ ineq2 ] ) that the entropy density for the gentile statistics majorizes that for the haldane - wu statistics .
such a proof will probably provide a deeper insight into general properties of exotic exclusion statistics .
also it would be interesting to verify the majorizing inequalities for the scaling functions ( [ crin ] ) and central charges ( [ cin ] ) related to the affine toda models and to understand whether such a majorization is a model independent feature of the two involved statistics .
i am grateful to a. fring for proposing the problem and participation in the initial stage of the research .
this work was supported in part by intas grant 99 - 01459 and by russian fund for fundamental investigations grant 99 - 01 - 00101 .
k. schoutens , _ phys .
* 79 * ( 1997 ) 2608 ; + a.g .
bytsko and a. fring , _ nucl .
_ * b521 * ( 1998 ) 573 , _ phys .
_ * b454 * ( 1999 ) 59 ; + p. bouwknegt and k. schoutens , _ nucl .
* b547 * ( 1999 ) 501 . a.j . bracken and h.s .
green , _ j. math .
* 14 * ( 1973 ) 1784 ; + p.g.o .
freund , _ phys .
* d13 * ( 1976 ) 2322 ; + m. cattani and n.c .
fernandes , _
nuovo cim .
_ * b87 * ( 1985 ) 70 ; _ phys .
lett . _ * a124 * ( 1987 ) 229 .
anand , s.n .
biswas and m. hasan , _ phys .
* d18 * ( 1978 ) 4529 ; + m.c . de sousa vieira and c. tsallis , preprint cbpf - nf-038/86 ( 1986 ) ; + p. suranyi , _ phys .
* 65 * ( 1990 ) 2329 ; + a.k .
rajagopal , _ phys .
_ * a214 * ( 1996 ) 127 .
d. bernard and y.s .
wu , cond - mat/9404025 ( 1994 ) ; + m.v.n .
murthy and r. shankar , _ phys .
* 73 * ( 1994 ) 3331 ; + k. hikami , _ phys
_ * a205 * ( 1995 ) 364 ; + s.b .
isakov , d.p .
arovas , j. myrheim and a.p .
polychronakos , _ phys .
_ * a212 * ( 1996 ) 299 .
fig . 2 : scaling functions of the @xmath293-minimal affine toda model for the gentile statistics of order @xmath302 ( solid curves , from down to up ) , for @xmath126 haldane - wu statistics ( the dashed curve ) and for the bosonic statistics ( the dotted curve ) .
fig . 3 : solution @xmath327 of the tba equation for the @xmath330-minimal affine toda model for the scaling parameter @xmath410 ( dotted curves ) , @xmath411 ( dashed curves ) , and @xmath412 ( solid curves ) .
for the same value of @xmath323 the lower curve corresponds to the fermionic statistics and the upper curve to @xmath224 gentile statistics .
fig . 4 : scaling functions of the @xmath330-minimal affine toda model for the gentile statistics of order @xmath413 ( solid curves , from down to up ) and for the haldane - wu statistics with @xmath414 ( the dashed curve ) .
fig . 5 : effective central charge of the @xmath293 and @xmath330-minimal affine toda models for the gentile statistics of order @xmath1 ( lower and upper solid curves ) and for the haldane - wu statistics with @xmath415 ( lower and upper dashed curves ) . | we investigate properties of the entropy density related to a generalized extensive statistics and derive the thermodynamic bethe ansatz equation for a system of relativistic particles obeying such a statistics .
we investigate the conformal limit of such a system .
we also derive a generalized y - system .
the gentile intermediate statistics and the statistics of @xmath0-ons are considered in detail . in particular
, we observe that certain thermodynamic quantities for the gentile statistics majorize those for the haldane - wu statistics .
specifically , for the effective central charges related to affine toda models we obtain nontrivial inequalities in terms of dilogarithms .
= november 2000 + hep - th/0011004 + 14.5pt |
the development of spin photonics has taken an important step forwards due to the recently experimental verifications of the spin hall effect ( she ) of light @xcite .
the she is a transport phenomenon , in which an applied field on the spin particles leads to a spin - dependent shift perpendicular to the electric field @xcite .
the she of light can be regarded as a direct optical analogy in which the spin electrons and electric potential are replaced by spin photons and refractive index gradient .
the she of light sometimes referred to as the fedorov
imbert effect , was predicted theoretically by fedorov @xcite , and was experimentally confirmed by imbert @xcite .
the theory was extended in a formulation which shows that the transverse spatial separation of the left and right circularly polarized components on oblique incidence directly from the total angular momentum conservation @xcite .
more recently , the interesting effect has also been observed in scattering from dielectric spheres @xcite , on the direction making an angle with the propagation axis @xcite , and in silicon via free - carrier absorption @xcite . the polarization - dependent transverse shift in the she of light is generally believed as a result of an effective spin - orbital interaction , which describes the mutual influence of the spin ( polarization ) and trajectory of the light beam @xcite .
there are two characteristics of the spin - orbit interaction of photons : the first one is the influence of the trajectory upon polarization @xcite ; the second one is the reciprocal influence of the polarization upon the trajectory @xcite .
these mechanisms have well been understood in gradient refractive index media @xcite .
in addition , the spin - orbital interaction in both inhomogeneous anisotropic media @xcite and in tightly focused beams @xcite can be explained by spin - to - orbital angular momentum conversion .
however , the physical picture of spin - orbit interaction in conventional beam refraction has not yet been fully examined .
for example , the relation between refractive index gradient and spin - to - orbital angular momentum conversion is unclear . whether spin - to - orbital angular momentum conversion can be enhanced ( or suppressed ) by increasing ( or decreasing ) the refractive index gradient ?
thus , the aim of this paper is to reveal the spin - to - orbital angular momentum conversion in the she of light from the viewpoint of classical electrodynamics .
the paper is organized as follows .
first , we establish a three dimension beam propagation model for describing the she of light in conventional beam refraction .
our result shows that the refractive index gradient can enhance or suppress the transverse spatial shifts in she of light .
we suggest that the metamaterial whose refractive index can be tailored arbitrarily is a good candidate to modulate the she of light .
next , we attempt to obtain a clear physical picture of spin - to - orbital angular momentum conversion in the she of light . within the paraxial approximation ,
a distinct separation between spin and orbital angular momenta is introduced .
we find that the she of light may be dramatically enhanced by facilitating the spin - to - orbital angular momentum conversion .
finally , we want to explore what role refractive - index gradient plays in the spin - to - orbital angular momentum conversion .
we demonstrate that the refractive index gradient can enhance or suppress the spin - to - orbital angular momentum conversion , and thus can control the she of light .
to reveal the spin - to - orbital angular momentum conversion , we need to establish a three dimension beam propagation model for describing the she of light .
figure [ fig1 ] illustrates the beam reflection and refraction in cartesian coordinate system .
the @xmath0 axis of the laboratory cartesian frame ( @xmath1 ) is normal to the air - glass interface locating at @xmath2 .
we use the coordinate frames ( @xmath3 ) for central wave vector , where @xmath4 denotes incident , reflection , and transmission , respectively .
in addition , we must introduce local cartesian frames ( @xmath5 ) to describe an arbitrary wave vector .
the electric field of the @xmath6th beam can be solved by employing the fourier transformations @xcite .
the complex amplitude for the @xmath6th beam can be conveniently expressed as @xmath7,\label{noapr}\end{aligned}\ ] ] where @xmath8 and @xmath9 is the angular spectrum . the approximate paraxial expression for the field in eq .
( [ noapr ] ) can be obtained by the expansion of the square root of @xmath10 to the first order @xcite , which yields @xmath11\label{apr}.\end{aligned}\ ] ] in general , an arbitrary linear polarization can be decomposed into horizontal and vertical components . in the spin basis
set , the angular spectrum can been written as : @xmath12 @xmath13 here , @xmath14 and @xmath15 represent horizontal and vertical polarizations , respectively .
the positive and negative signs denote the left and right components , respectively @xcite .
the monochromatic gaussian beam can be formulated as a localized wave packet whose spectrum arbitrarily narrows , whose angular spectrum can be written as @xmath16\label{asi},\ ] ] where @xmath17 is the beam waist .
after the angular spectrum is known , we can obtain the field characteristics for the @xmath6th beam .
( color online ) schematic illustrating the reflection and refraction of central and local wave vectors at an air - glass interface in cartesian coordinate system .
@xmath18 are are reference frames for central wave vector and @xmath19 are local reference frames for an arbitrary wave vector . @xmath20 and @xmath21 are the interface reference frames for central and local wave vectors , respectively.,width=302 ] to accurately describe the she of light , it is need to determine the reflection and transmission of arbitrary wave - vector components . from the central frame @xmath22 to the local frame @xmath23 , the following three steps should be carried out .
first , we transform the electric field from the reference frame @xmath22 around the @xmath24 axis by the incident angle @xmath25 to the frame @xmath26 : @xmath27 , where @xmath28.\label{matrixr}\end{aligned}\ ] ] then , we transform the electric field from the reference frame @xmath26 around the @xmath24 axis by an angle @xmath29 to the frame @xmath30 , and the correspondingly matrix is given by @xmath31.\label{matrixr}\end{aligned}\ ] ] finally , we transform the electric field from the reference frame @xmath30 around the @xmath24 axis by an angle @xmath32 to the frame @xmath23 , and the rotation matrix can be written as @xmath33.\label{matrixr}\end{aligned}\ ] ] the rotation matrix of the reference frame @xmath22 around the @xmath24 axis by an angle @xmath34 , and we have @xmath35.\label{matrixr}\end{aligned}\ ] ] for an arbitrary wave vector , the reflected field is determined by @xmath36 , where @xmath37 and @xmath38 are the fresnel reflection coefficients .
the reflected field should be transformed from @xmath39 to @xmath40 . following the similar procedure ,
the reflected field can be obtained by carrying out three steps of transformation : @xmath41 where @xmath42.\label{matrixr}\end{aligned}\ ] ] here , only the two dimension rotation matrices is taken into account , since the longitudinal component of electric field can be obtained from the divergence equation @xmath43 .
the reflection matrix can be written as @xmath44m_{{x_iy_iz_i}\rightarrow{x_iy_iz_i}}.\label{matrixri}\end{aligned}\ ] ] the reflected angular spectrum is related to the boundary distribution of the electric field by means of the relation @xmath45 , and we have @xmath46 = \left [ \begin{array}{cc } r_p&\frac{k_{ry } ( r_p+r_s ) \cot\theta_i}{k_0 } \\
-\frac{k_{ry } ( r_p+r_s)\cot\theta_i}{k_0 } & r_s \end{array } \right]\left[\begin{array}{cc } \tilde{e}_i^h\\ \tilde{e}_i^v \end{array}\right],\label{matrixrii}\end{aligned}\ ] ] it is well known that @xmath37 and @xmath38 can be expanded as a power series that can be truncated to an appropriate order @xmath47 . by making use of taylor series expansion based on the arbitrary angular spectrum component , @xmath37 and @xmath38
can be expanded as a polynomial of @xmath48 : @xmath49_{k_{ix}=0}\nonumber\\ & & + \sum_{j=2}^{n}\frac{k_{ix}^n}{j!}\left[\frac{\partial^j r_{p , s}(k_{ix})}{\partial k_{ix}^j}\right]_{k_{ix}=0}\label{lmd}.\end{aligned}\ ] ] our analysis is confined to the zero order to obtain a sufficiently good approximation . from the boundary condition , we obtain @xmath50 and @xmath51 .
in fact , after the incident angular spectrum is known , eq .
( [ apr ] ) together with eqs .
( [ asi ] ) and ( [ matrixrii ] ) provides the paraxial expression of the reflected field : @xmath52\label{hpr}.\end{aligned}\ ] ] @xmath53\label{vpr}.\end{aligned}\ ] ] where @xmath54 and @xmath55 .
note that the above expression of reflected field coincides with the early results @xcite with different methods .
we proceed to consider the transmitted field . following the similar procedure
, we obtain the transform matrix from @xmath56 to @xmath57 as @xmath58,\label{matrixr}\end{aligned}\ ] ] where @xmath59 is the transmitted angle . for an arbitrary wave vector , the transmitted field
is determined by @xmath60 , where @xmath61 and @xmath62 are the fresnel transmission coefficients .
hence , the transmitted field should be transformed from @xmath39 to @xmath57 , and the transmission matrix can be written as @xmath63m_{{x_iy_iz_i}\rightarrow{x_iy_iz_i}}.\label{matrixti}\end{aligned}\ ] ] the transmitted angular spectrum is related to the boundary distribution of the electric field by means of the relation @xmath64 , and can be written as @xmath65 = \left [ \begin{array}{cc } t_p&\frac{k_{ty } ( t_p-\eta t_s ) \cot\theta_i}{k_0 } \\ \frac{k_{ty } ( \eta t_p - t_s)\cot\theta_i}{k_0 } & t_s \end{array } \right]\left[\begin{array}{cc } \tilde{e}_i^h\\ \tilde{e}_i^v \end{array}\right],\label{matrixtii}\end{aligned}\ ] ] where @xmath66 .
based on the taylor series expansion , the fresnel transmission coefficients @xmath61 and @xmath62 can be written as @xmath67_{k_{ix}=0}\nonumber\\ & & + \sum_{j=2}^{n}\frac{k_{ix}^n}{j!}\left[\frac{\partial^j t_{p , s}(k_{ix})}{\partial k_{ix}^j}\right]_{k_{ix}=0}\label{tps}.\end{aligned}\ ] ] note that the fresnel coefficients are real in the regime of partial reflection and transmission , @xmath68 . from the snell s law under the paraxial approximation , we obtain @xmath69 and @xmath70 . substituting eqs .
( [ asi ] ) and ( [ matrixtii ] ) into eq .
( [ apr ] ) , we obtain the transmitted field : @xmath71\label{hpt}.\end{aligned}\ ] ] @xmath72\label{vpt}.\end{aligned}\ ] ] here , @xmath73 and @xmath74 .
the interesting point we want to stress is that there are two different rayleigh lengths , @xmath75 and @xmath76 , characterizing the spreading of the beam in the direction of @xmath77 and @xmath24 axes , respectively . up to now , we have established a general propagation model to describe the reflected and transmitted fields .
it is well known that the she of light manifests itself as polarization - dependent transverse shifts in the process of reflection and refraction . to reveal the she of light
, we now determine the transverse spatial shifts of field centroid . the time - averaged linear momentum density associated with the electromagnetic field
can be shown to be @xcite @xmath78\label{lmd},\ ] ] where the magnetic field can be obtained by @xmath79 .
the intensity distribution of electromagnetic fields is closely linked to the longitudinal momentum currents @xmath80 .
( color online ) the she of light manifests itself as polarization - dependent transverse shifts of field centroid . [ ( a ) , ( b ) ] intensity distribution of the reflected field for left and right circularly polarized components , respectively .
[ ( c ) , ( d ) ] intensity distribution of the transmitted field for left and right circularly polarized components , respectively .
the refractive index of the glass is @xmath81 and the incident angle is chosen as @xmath82 .
the beam waist is chosen as @xmath83 .
the intensity distributions in the plane @xmath84 are plotted in normalized units.,width=302 ] figure [ fig2 ] shows the transverse spatial shifts of beam centroid in an air - glass interface . in the case of reflection
, the left circularly polarized component exhibits a negative shift [ fig .
[ fig2](a ) ] . for the right circular polarization , however , presents a positive shift [ fig .
[ fig2](b ) ] . in the case of transmission ,
the left circularly polarized component exhibits a positive transverse shift [ fig .
[ fig2](c ) ] . for the right circularly polarized component
, however , presents a negative transverse shift [ fig .
[ fig2](d ) ] .
this interesting phenomenon gives a clear evidence of polarization - dependent splitting of field intensity in the she of light .
the transverse shifts are polarization - dependent , and thus can be regarded as the influence of the polarization upon trajectory . in the air - glass interface ,
the transverse shifts are just a few tens of nanometers , which can be observed via weak measurements @xcite .
however , how to amplify this tiny effect is still an open problem . as we know that the refractive index gradient acts as the electric potential gradient in the electronic systems .
now a question naturally arises : whether can the refractive index gradient enhance the transverse shift in the she of light ? to answer this question we need to obtain a relation between the transverse shift and the refractive index gradient . at
any given plane @xmath85 , the transverse spatial shift of beam centroid compared to the geometrical - optics prediction is given by @xmath86 note that the transverse spatial shift is @xmath87-independent , while the transverse angular shift can be regarded as a small shift inclining from the @xmath87 axis .
in addition , the field also experience a longitudinal spatial shift @xcite and a longitudinal angular shift @xcite .
note that the angle shift means that the snell s law can not accurately describe the beam refraction phenomenon @xcite .
we first consider the transverse spatial shift of the reflected field . after substituting the reflected field eqs .
( [ hpr ] ) and ( [ vpr ] ) into eq .
( [ centroid ] ) , we obtain the transverse spatial shifts as @xmath88 @xmath89 for an arbitrary linearly polarized beam , the transverse spatial shift of the reflected field is given by @xmath90 where @xmath91 is the reflected polarization angle . in the frame of classical electrodynamics , the reflection polarization angle is determined by : @xmath92 @xmath93 here , @xmath94 is the incident polarization angle . under the limit of ultra - high refractive index gradient
, the transverse spatial shifts can be written as @xmath95 .
this simple result means that the recent advent of a new class of metamaterial with ultra - high refractive index @xcite is a possible candidate to eliminate the transverse spatial shift in the reflected field .
( a ) schematically shows the she of light manifests itself as spin - dependent splitting .
the transverse spatial shifts of transmitted field @xmath96 versus the thickness of air gap @xmath97 .
the incident polarization angles are chooser as ( b ) @xmath98 , ( c ) @xmath99 , and @xmath100 .
other parameters are chosen to be the same as in fig .
[ fig2].,width=302 ] we next consider the transverse spatial shifts of the transmitted field . after substituting the transmitted field eqs .
( [ hpt ] ) and ( [ vpt ] ) into eq .
( [ centroid ] ) , we have @xmath101 @xmath102 for an arbitrary linearly polarized wave - packet , the transverse spatial shifts of the transmitted field are given by @xmath103 where the transmission polarization angle @xmath104 is determined by @xmath105 @xmath106 note that the above expression coincides well with the early results @xcite with the quantum method .
our scheme shows that the she of light can be explained from the classic electrodynamics . under the limit of ultra - high refractive index gradient
, the transmitted field tends to reach a saturation value : @xmath107 up to now , we have described the polarization - dependent splitting in the she of light from the viewpoint of pure classical electrodynamics .
we proceed to examine the role of refractive index gradient ( i.e. , @xmath108 ) in the she of light .
the normalized transverse spatial shifts of transmitted field for various refractive index gradients as shown in fig .
we first consider the wave packet incident from air to a low - refractive - index medium ( @xmath109 ) .
for the left circularly polarized component , the field centroid exhibits a negative transverse shift .
for the right circularly polarized component , the beam centroid also presents a transverse spatial shift , but in an opposite direction .
we next consider the wave packet incident from air to a high - refractive - index medium ( @xmath110 ) .
for a certain polarized component , we find that the she of light is reversed when the refractive index gradient is inverted .
it is clearly shown that the transverse spatial shifts increases with the increase of the refractive index gradient @xmath111 .
within the regime of low refractive index , the transverse spatial shifts are enhanced sharply with the increase of @xmath111 .
when the refractive index @xmath112 , the beam is totally reflected . while in the regime of high refractive index , the transverse spatial shifts tend to reach a saturation value .
in addition , the transverse spatial shifts are also sensitive to the incident polarization angle as clearly shown in the figure .
hence , we can enhance or suppress the transverse spatial shifts in she of light by modulating the refractive index gradient and incident polarization angle .
the normalized transverse spatial shifts of transmitted field @xmath113 versus incident angles @xmath25 in the metamaterial with @xmath114 and @xmath115 ( dashed - dotted and dashed - doted - dotted lines ) . as a comparison
, the conventional medium is chosen as the glass with @xmath116 ( solid and dashed lines ) . the incident polarization angle
choose as @xmath117 .
other parameters are chosen to be the same as in fig .
inset : zoomed - in view of the transverse spatial shifts at the air - metamaterial interface .
, width=302 ] how to attenuate the she of light is not of pure academic interest , owing to the requirement of eliminating the transverse shifts , such as reflection , refraction , and focusing in optical experiments .
we suggest that the metamaterial whose refractive index can be tailored arbitrarily @xcite may become a good candidate . in order to accurately describe how the metamaterial attenuate the she of light , it is necessary to include material dispersion and losses .
thus , a certain dispersion relation , such as the lorentz medium model , should be introduced .
the relative constitutive parameters are @xmath118 to avoid the trouble involving in a certain value of frequency , we assume the material parameters are @xmath119 , @xmath120 , @xmath121 , and @xmath122 .
figure [ fig4 ] shows the normalized transverse spatial shifts of transmitted field versus incident angle at the air - metamaterial interface . for the comparison
, we also plot the transverse spatial shifts at the air - glass interface . in a large range of incident angles ,
the metamaterial is a good candidate for suppressing the she of light in the process of refraction . in the ideal case ( @xmath123 and @xmath124 )
, the transverse shift can be eliminated completely .
this is another reason why a simple planar slab provides us with a perfect lens without aberration @xcite .
as we will see in the following , the spin - to - orbital angular momentum conversion can be used to explain the inherent physics underlying these intriguing phenomena .
it should be noted is that the spin - orbital interaction in both inhomogeneous anisotropic media @xcite and in tightly focused beams @xcite can be explained by spin - to - orbital angular momentum conversion . to obtain a clear physical picture of the she of light
, we introduce a distinct separation between spin and orbital angular momenta to clarify the spin - orbital interaction .
the momentum current can be regarded as the combined contributions of spin and orbital parts : @xmath125 here , the orbital term is determined by the macroscopic energy current with respect to an arbitrary reference point and does not depend on the polarization .
the spin term , on the other hand , relates to the phase between orthogonal field components and is completely determined by the state of polarization @xcite . in a monochromatic optical beam , the spin and orbital currents can be respectively written in the form : @xmath126,\ ] ] @xmath127,\ ] ] where @xmath128 is the invariant berry notation @xcite .
it has been shown that both spin and orbital currents originate from the beam transverse inhomogeneity and their components are directly related to the azimuthal and radial derivatives of the beam profile parameters .
the orbital currents are mainly produced by the phase gradient , while the spin currents are orthogonal to the intensity gradient .
( color online ) the transverse angular momentum density of left circularly polarized component .
[ ( a ) and ( b ) ] incident fields : @xmath129 and @xmath130 . [ ( c ) and ( d ) ] transmitted fields : @xmath131 and @xmath132 .
the cross section is chosen as @xmath133 and the intensity is plotted in normalized units .
other parameters are the same as those in fig .
[ fig2].,width=302 ] we proceed to analysis the angular momentum density for each of individual wave packet , which can be written as @xmath134 within the paraxial approximation , the angular momentum can be divided into the orbital and spin angular parts : @xmath135 @xcite , it follows that @xmath136 @xmath137 it should be mentioned that this separation still hold beyond the paraxial approximation @xcite .
the transverse angular momentum density @xmath138 can be regarded as the combined contributions of orbital and spin parts : @xmath139 @xmath140 the time - averaged linear and angular momenta can be obtained by integrating over the whole @xmath141 plane @xcite @xmath142 @xmath143 the transverse orbital angular momentum would appear if the spin - to - orbital angular momentum conversion occurs in the process of reflection and refraction .
it should be mentioned that the transverse orbital angular momentum can lead to a transverse shift of field centroid even in free space @xcite . in the plane
@xmath84 , after substituting eqs ( [ oamdx ] ) and ( [ samdx ] ) into eqs ( [ oam ] ) and ( [ sam ] ) we obtain @xmath144 figure [ fig5 ] plots the transverse spin and orbital momentum currents for incident and transmitted fields .
it is clearly shows that the transverse orbital angular momentum plays a dominant role in the she of light .
compare fig [ fig5](b ) with ( d ) , however , it is not very clear why the transmitted field exhibits the she of light .
thus , it is necessary to determine the transverse orbital angular momentum . from eq .
( [ oam ] ) , we can obtain @xmath145 and @xmath146 .
the presence of transverse orbital angular momentum @xmath147 provide a direct evidence for the spin - to - orbital angular momentum conversion . in the she of light associated with refraction @xcite , reflection @xcite , and tunneling @xcite , the transverse spatial shifts are sensitive to the incident polarization angles .
we proceed to explore the role of refractive index gradient in spin - to - orbital angular momentum conversion
. the @xmath0 component of total angular momentum @xmath148 for the @xmath6th beam can be represented as a sum of the extrinsic orbital angular momentum @xmath149 and the intrinsic spin angular momentum @xmath150 , i.e. , @xmath151 @xcite .
the @xmath0 component of the orbital angular momenta are given by @xmath152 , @xmath153 , and @xmath154 .
for incident , reflected , and transmitted wave packets , respectively . in the laboratory cartesian frame ( @xmath1 ) , the @xmath0 component of the orbital angular momenta are given by @xmath155 @xmath156 @xmath157 the @xmath87 components of spin angular momentum for the @xmath6th beam is respectively described by @xmath158 @xmath159 @xmath160 from eqs .
( [ jizo])-([jtzo ] ) and eqs .
( [ jizs])-([jtzs ] ) , we find that the angular momenta fulfill the conservation law : @xmath161 here , @xmath162 and @xmath163 are the energy reflection and energy transmission coefficients , respectively .
note that the transverse angular shifts , which governed by the linear momentum conservation law , are not discussed in this paper . the refractive index gradient @xmath164 induces the spin - to - orbital angular momentum conversion .
the incident wave packets are chosen as left circular polarization and right circular polarization .
other parameters are chosen to be the same as in fig .
[ fig2].,width=302 ] to obtain a clear physical picture of the she of light , the spin - to - orbital angular momentum conversion induced by refractive index gradient is depicted in fig .
[ fig6 ] . for
the left circularly polarized component , the spin angular momentum @xmath165 monotonously increase while the orbital angular momentum @xmath166 monotonously decreases with the increase of the refractive index gradient @xmath111 .
for the right circular polarization , both the spin angular momentum @xmath167 and the orbital one @xmath168 present oppositive features . when the refractive index gradient continues increasing , the spin - to - orbital angular momentum conversion appears to reach a saturation value : @xmath169 this gives a good explanation of the intriguing phenomena why the transverse spatial shifts tends to saturation values in the fig .
the metamaterial with @xmath170 is a good candidate to eliminate the she of light in the refraction , since the spin - to - orbital angular momentum conversion is impressed completely .
the spin - to - orbital angular momentum conversion can be enhanced in the region of low refractive index gradient as shown in fig .
[ fig6 ] . as a result ,
the transverse spatial shifts are also be amplified in this region . however , when the refractive index @xmath112 the wave pack is totally reflected .
by properly facilitating the spin - to - orbital angular momentum conversion the she may be enhanced dramatically .
we now give a very simple way to understand how the refractive index gradient enhance the spin - to - orbital angular momentum conversion in the she of light .
we attempt to perform analyses on the @xmath0 component of the total angular momentum for each of individual photons , i.e. , @xmath171 @xcite .
the total angular momentum conservation law for single photon is given by @xmath172 where @xmath173 .
when the photons penetrate from air into a low - refractive - index medium ( @xmath174 ) , the incident angle is less than the transmitted angle @xmath175 . for the left circularly polarized photons , the @xmath176 component of spin angular momentum @xmath177 decreases after entering the medium . because of the conservation law
, the total angular momentum must remain unchanged . to conserve the total angular momentum
, the photons must move to the direction @xmath178 ( @xmath179 ) and thus generate a positive orbital angular momentum ( @xmath180 ) .
for the right circularly polarized photons , the @xmath0 component of spin angular momentum @xmath181 increases . in this case
, the photons must move to the direction @xmath182 ( @xmath183 ) and induce a negative orbital angular momentum ( @xmath184 ) . when the photons enter into a high - refractive - index medium ( @xmath185 ) , the incident angle is larger than the transmitted angle @xmath186 . as a result ,
the orbital angular momentum reverses its direction .
in conclusion , we have identified the role of spin - to - orbital angular momentum conversion in she of light .
we have demonstrated that the refractive index gradient can enhance or suppress the spin - to - orbital angular momentum conversion , and thus can control the she of light .
the recent advent of metamaterial whose refractive index can be tailored arbitrarily seems to be an available candidate to amplify or eliminate the she of light . however , the spin - to - orbital angular momentum conversion in a ultra - large refractive index gradient is limited by a saturation value .
fortunately , the she of light can be dramatically amplified by plasmonic nanostructure @xcite . in addition , the she of light can also be noticeably enhanced when the beam carries orbital angular momentum @xcite .
hence , the exploration of spin - to - orbital angular momentum conversion in these systems would be very interesting .
the transverse spatial shifts governed by the spin - to - orbital angular momentum conversion , give us a clear physical picture to clarify the role of refractive index gradient in the she of light .
these findings provide a pathway for modulating the she of light , and thereby open the possibility for developing new nano - photonic devices . because of the close similarity in optical physics @xcite , condensed matter @xcite , and high - energy physics @xcite , by properly facilitating the spin - to - orbital angular momentum conversion ,
the she may be enhanced dramatically in these physical systems .
one of authors ( h.l . )
thanks dr .
w. lffler for helpful discussions .
this research was partially supported by the national natural science foundation of china ( grants nos .
61025024 , 11074068 , and 10904036 ) . | from the viewpoint of classical electrodynamics , we identify the role of spin - to - orbital angular momentum conversion in spin hall effect ( she ) of light .
we introduce a distinct separation between spin and orbital angular momenta to clarify the spin - orbital interaction in conventional beam refraction .
we demonstrate that the refractive index gradient can enhance or suppress the spin - to - orbital angular momentum conversion , and thus can control the she of light .
we suggest that the metamaterial whose refractive index can be tailored arbitrarily may become a good candidate for amplifying or eliminating the she of light , and by properly facilitating the spin - to - orbital angular momentum conversion the she may be enhanced dramatically . the transverse spatial shifts governed by the spin - to - orbital angular momentum conversion , provide us a clear physical picture to clarify the role of refractive index gradient in the she of light .
these findings provide a pathway for modulating the she of light and can be extrapolated to other physical systems . |
globular clusters are the fossil record of galaxy formation .
they are among the oldest known objects in the universe .
since the first colour - magnitude diagrams ( cmd ) for gcs were obtained by arp , baum and sandage ( 1952 ) in the early 50 s , gcs became the natural laboratory where the theory of stellar evolution was tested .
on the other hand , the stellar evolution theory has allowed the determination of the age of gcs ( e.g. , sandage 1962 , iben & renzini 1984 , vandenberg et al .
1992 , jimenez et al .
1996 ) , that is today considered as the best estimate for the age of the universe .
each galaxy contains hundreds of gcs , whose properties are surprisingly similar in the universe ( harris and racine 1979 , harris 1991 ) , suggesting the presence of a common physical mechanism in the early stages of galaxy formation .
the study of gcs should therefore provide important clues for the development of a theory for the origin of galaxies .
sites of present day star formation do not form bound stellar systems as massive as gcs .
therefore modelling the formation of gcs could be a critical test for any theory of star formation .
it is known that stars are formed with an efficiency of a few percent in giant molecular clouds of the galactic disk ( duerr , imhoff & lada 1982 , myers et al .
1986 , mooney and solomon 1988 ) , and bound open clusters with an efficiency ten times smaller ( larson 1986 ) .
therefore it is not surprising if gcs contain only 0.1% of the luminous mass of the parent galaxy , since , from the point of view of the efficiency of present day star formation , they can only emerge from protoglobular clouds 1000 times more massive ( as in searle 1977 , searle & zinn 1978 , harris & pudritz 1994 )
. nevertheless models of gcs formation have identified objects of just @xmath7 m@xmath2 with protoglobular clouds .
such models place the gc formation period either before the galaxy is formed or during its formation . the first model for gcs as primordial objects was proposed by peebles & dicke 1968 , and was later revised by peebles ( 1984 ) and rosenblatt , faber & blumenhal ( 1988 ) .
the secondary formation scenario , in which the mass of the protoglobular clouds are determined by the detailed mechanism of cooling in the protogalactic cloud , was first proposed by fall & rees ( 1985 ) and later improved by kang et al .
( 1990 ) , ashman ( 1990 ) , brown , burkert & truran ( 1991 ) , brown , burkert & truran ( 1995 ) , vietri & pesce ( 1995 ) .
an alternative scenario has been proposed to explain the formation of some of the gcs in large shocks , e.g. in merging and interacting galaxies ( ashman & zepf ( 1992 ) , kumai , basu & fujimoto ( 1993 ) ) . in this paper
we explore the possibility that gcs are formed in large clouds identified with primordial mass fluctuations in a cold dark matter ( cdm ) spectrum .
the progenitors of gcs are clouds as massive as a few @xmath0 m@xmath2 of baryons ( searle & zinn 1978 , zinnecker et al .
1988 and larson 1990 have suggested that gcs form in the core of very massive clouds ) .
such mass - scale is selected out of the standard cdm spectrum by the mechanism of non - equilibrium h@xmath8 formation ( see section 2 ) and , as mentioned above , it is required by the low efficiency of the process of star formation on large scales in order to form a bound stellar system .
the protoglobular cloud is efficiently cooled to approximately 100 k by h@xmath8 collisional excitation .
the cooling time is short enough that the gas can be considered isothermal . during the dissipative collapse
( isothermal shocks radiate away most of the kinetic energy ) of the baryonic gas ( see section 2 ) star formation occurs until supernova explosions disperse away the gas ( section 4 ) .
we suggest that this star formation process in very massive primordial clouds can be responsible for the formation of the halo gcs ( gcs with metallicity below 0.1 the solar value ) and of a significant fraction of the halo stars .
the dispersed metal - enriched gas is further processed in following star formation episodes .
the paper is organised in the following way . in section 2
we show how the mass - scale of protoglobular clouds is selected .
section 3 is a brief review of the statistical model of star formation ( padoan 1995 ) that is used in this work .
results are presented in sections 4 and 5 .
the last two sections of the paper contain the discussion and the conclusions .
if gcs have a primordial origin , then the mass of the typical protoglobular cloud should be of the order of @xmath9 .
on the other hand , it is observed that the specific frequency of gcs ( harris & racine ( 1979 ) ) in different spiral galaxies is approximately constant , which means that there must be a physical mechanism responsible for the value of the mass of the primordial protoglobular cloud .
the value inferred from this argument is of a few @xmath0 m@xmath2 .
the rms fluctuation @xmath10 in the mass inside a sphere of a given radius , which is an integral over the power spectrum of density fluctuations , was determined by peblees ( 1984 ) for a cold dark matter ( cdm ) scenario .
@xmath10 is a shallow function of the mass scale without any particular feature . nevertheless , a mass scale for the first baryonic fluctuations that collapse can be identified .
baryonic fluctuations of mass smaller than this will not collapse and fragment earlier .
the mass - scale is selected out of the cdm spectrum as a consequence of the mechanism of non - equilibrium formation of h@xmath8 .
in fact the collapse and fragmentation of the gas in a gravitational potential dominated by collisionless dark matter particles requires energy dissipation , which becomes efficient as soon as a strong cooling mechanism is activated .
we recognise such cooling mechanism in the @xmath3 collisional excitation .
the mass - scale arises because the formation of @xmath3 in the primordial gas ( peebles & dicke 1968 , hirasawa 1969 , hirasawa , aizu & taketani 1969 , matsuda , sato & takeda 1969 , hutchins 1976 , silk 1977 , carlberg 1981 , palla , salpeter & stahler 1983 , lepp & shull 1984 ) occurs during the non - equilibrium recombination ( cooling time shorter than recombination time ) of the h@xmath8 , that is in gas after strong ionising shocks ( lepp & shull 1983 , dove & mandy 1986 , shapiro & kang 1987 , palla & zinnecker 1988 ) . the velocities necessary to produce such strong shocks are not present during the collapse of clouds of mass smaller than @xmath0 m@xmath2 .
therefore efficient energy dissipation occurs only in primordial clouds of baryonic mass larger than @xmath0 m@xmath2 , and these clouds will be the first , in the cdm bottom - up hierarchy , to experience the fragmentation and collapse of their baryonic gas .
the mass of the clouds given by this argument is comparable to the mass of protoglobular clouds as inferred above from the approximately constant specific frequency of gcs in galaxies .
we identify these primordial clouds of a few @xmath0 m@xmath2 with protoglobular clouds .
a new statistical model of star formation has been recently discussed in a series of papers ( padoan 1995 , nordlund & padoan 1996 , padoan , jones , & nordlund 1996 , padoan , nordlund , & jones 1996 ) .
the model is relevant when the star forming gas is characterized by : * random supersonic motions ; * cooling time shorter than dynamical time of the random motions . under these conditions
a complex system of interacting isothermal shocks is present in the gas .
such flow creates a highly nonlinear density field , with density contrasts of a few order of magnitudes .
since random motions are probably ubiquitous in sites of star formation , we suggest to describe the formation of protostars as the gravitational collapse of jeans masses in a density distribution shaped by random supersonic motions . we therefore identify a protostar with one jeans mass .
the advantage of this description of star formation is that a statistical description of the density field arising from random motions in the gas can be given , and it does not depend on the detailed phisics , but just on the rms mach number of the flow . in the following subsections we briefly present the results of recent numerical simulations of supersonic randomly forced flows , their observational counterpart , and the derivation of the protostar mass function ( mf ) based on the numerical ( and observational ) results .
nordlund and padoan ( 1996 ) and padoan , nordlund and jones ( 1996 ) have recently discussed the importance of supersonic flows in shaping the density distribution in the cold interstellar medium ( ism )
. they have run numerical simulations of isothermal flows randomly forced to high mach numbers
. a description of the code and the details of the experiments can be found in nordlund & padoan ( 1996 ) .
here we just mention that the experiments consist in sloving the compressible mhd equations in a 3-d periodic mesh .
a random force is applied in fourier space to small wave numbers .
the random force produces a random flow , which is supersonic ( rms mach number up to 10 ) .
an isothermal equation of state is used , so that the flow soon develops a complex system of interacting shocks , with density contrasts of a few orders of magnitudes .
it is found that most of the mass concentrates in a small fraction of the total volume of the simulation , with a very intermittent distribution .
the probability density function ( pdf ) of the density field is well approximated by a log - normal distribution : @xmath11dlnx \label{1}\ ] ] where @xmath12 is the relative number density : @xmath13 and the standard deviation @xmath14 and the mean @xmath15 are functions of the rms mach number of the flow , @xmath16 : @xmath17 and @xmath18 or , for the linear density : @xmath19 where @xmath20
. therefore the standard deviation grows linearly with the rms mach number of the flow .
it is also found that the power spectrum , @xmath21 , of the density distribution is consistent with a power law : @xmath22 where @xmath23 is the wavenumber .
an observational counterpart of these numerical results have been recently recognized by padoan , jones , & nordlund ( 1996 ) .
they show that the infrared stellar extinction data presented by lada et al .
( 1994 ) can be easily interpreted if the density distribution of the absorbing dark cloud is very intermittent .
the observations are consistent with a log - normal distribution .
padoan , jones , & nordlund have shown that the standard deviation and the power spectrum of the 3-d density distribution in the dark cloud can be constrained by the stellar extinction measurements .
they find values of standard deviation and spectral index that are consistent with the numerical prediction . in the following subsection
, we show how to derive the mf of protostars , using the random density field predicted numerically and confirmed observationally in dark clouds .
a simple way to derive the protostar mf is that of defining a protostar as one local jeans mass , so that the protostar mf is simply a jeans mass distribution .
the jeans mass distribution is then just determined by the density distribution , because the gas is cooling rapidly and therefore the temperature is uniform .
in our scenario , random supersonic motions ( cascading from larger scale ) are present , and are responsible for shaping the density field .
strong density enhancements are due to the convergence of the flow , that is due to nonlinear hydrodynamical interactions , rather than to the local gravitational potential .
we suggest therefore a description of star formation where random motions are first creating a complex and highly nonlinear density field ( through isothermal shocks ) , and gravity then takes over , when each ` local ' jeans mass ( defined with the local density ) collapses into a protostar .
let us start from equation ( [ 1 ] ) , that is the statistic of the density , that means the fraction of the volume occupied by any given value of the density .
if we multiply that function times the relative density @xmath12 , we get the fraction of the _ mass _ occupied by any given density : @xmath24 now let us imagine to filter the density field with a filter of radius @xmath25 , that corrisponds to the mass scale @xmath26 .
we can sample the density distribution using this filter , and repeat the operation for different filters . for any scale @xmath26
we get a distribution of density @xmath27 . we know
, from the numerical simulations and from the observations mentioned above , that the log - normal distribution is a good model of the statistic of the density field
. therefore we have a log - normal distribution , @xmath28 , for every mass scale @xmath26 , each with its own value for the standard deviation , @xmath29 , and for the mean , @xmath30 .
the fraction of the total mass in collapsing structures of scale @xmath26 is the integral of the distribution @xmath28 along relative densities @xmath31 : @xmath32 where @xmath33 is the jeans density for the mass @xmath26 .
but of course the structure of radius @xmath25 can contain many local jeans masses , especially if it is very dense , and therefore can fragment into smaller objects .
for this reason we must repeat the same argument for a mass scale @xmath34 , and subtract the collapsing mass fraction in objects of radius @xmath35 , from the previously calculated collapsing mass in objects of radius @xmath25 : @xmath36 the situation is illustrated in fig.1a , where two filter masses , @xmath37 and @xmath38 are shown .
now , one should go to the limits @xmath39 and @xmath40 , and take the derivative along mass of an integral along density , @xmath41 where also the lower extreme of integration is subject to the derivative .
the whole procedure of extracting the protostar mf from the density distribution can be made much easier , by assuming the distribution @xmath28 to be the same for any @xmath26 .
this is in fact a good approximation , because we know from the numerical experiments and from the observations mentioned above , that the power spectrum is @xmath42 , where the spectral index @xmath43 .
the standard deviation @xmath29 is the mass variance that defines the power spectrum : @xmath44 from which we can write : @xmath45 where @xmath46 is the spectral index . for a spectral index @xmath47
we obtain @xmath48 .
since the exponent is so small , the distribution of density is approximately the same for any filter :
@xmath49 in other words we can say that the power spectrum is steep enough that the density field is almost self - similar , when looked through different filters .
fig.1b is the simplified version of fig.1a . now
the problem of deriving the mass function has become trivial , since the subtraction of the collapsing structures of scale @xmath50 , from the probability of collapse at the scale @xmath26 , is simply the derivative , along masses , of an integral along density , whose integrand do not depend on mass : @xmath51 the solution is just the value of the integrand at @xmath52 , times @xmath53 .
so we can see that the mf for the protstars , @xmath54 , is just given by the transformation of the density distribution into a jeans mass distribution : @xmath55 where @xmath56 is the jeans mass for the relative density @xmath12 , or vice versa .
the jeans mass can be written as : @xmath57 where : @xmath58 is the average jeans mass , that is the jeans mass for the average density @xmath59 .
here we use the simplest definition of jeans mass : without turbulent pressure or rotation , because the gas has just been shocked and is dissipating its kinetic energy in a short time ; without magnetic pressure , because we will discuss the role of the magnetic field in such random flows in subsequent papers ( our numerical experiments are in fact solving the mhd equations ) .
using equations ( [ 1 ] ) , ( [ 7 ] ) , ( [ 10 ] ) , ( [ 11 ] ) , and ( [ 12 ] ) we get the protostar mf : @xmath60 dm \label{13}\ ] ] where @xmath56 is in solar masses , and : @xmath61 one can also express the mf in average jeans mass , instead of in solar masses : @xmath62d(\frac{m}{b})\ ] ] a log - log plot of the protostar mf is shown in fig.2 .
one can recognize a long tail at large masses and an exponential cutoff at the smallest masses , inherited from the log - normal distribution of density .
this shape is a good result , because most models for the origin of the stellar imf are not able to reproduce the cutoff at the smallest masses , which should be present in any reasonable imf .
the dependence of the mf on the physical parameters of the star forming gas ( average temperature , density and velocity dispersion ) is discussed in padoan , nordlund , & jones ( 1996 ) , where a comparison with the observations is also presented .
if the mf is expressed per linear mass interval ( eg number of stars rather than mass fraction ) , its maximum occurs at the mass : @xmath63 ( padoan , nordlund , jones 1996 ) .
this is the typical stellar mass .
the result of the present derivation of the mf is that basically all the gas turns into stars ( of different masses ) , after the density distribution ( [ 1 ] ) is established .
this is a process that takes about one dynamical time of the random motions on the large scale .
therefore the star formation efficiency , defined as the star formation rate per unit mass , is just the inverse of the dynamical time of the random motions .
since the typical mass of the protoglobular cloud is determined by the condition that @xmath3 can be formed during the collapse , the cooling time in the gas ( @xmath64 yr at a temperature of 200 - 300 k ) is much shorter than the free - fall time ( a few @xmath0 yr ) . therefore the cloud is collapsing at a constant temperature of about 100 k. the collapse will be a turbulent flow with a complex system of strong isothermal shocks , where high density fluctuations are produced and protostars are born according to the model of turbulent fragmentation . the high reynolds number in the gas flows , the initial density inhomogeneities expected from the cdm scenario and the shear motions due to the tidal field are together responsible for the turbulence during the collapse .
moreover , any initial turbulence will be exponentially amplified during the collapse ( rotman 1991 , jacquin , cambon & blin 1993 ) . in order to apply the statistical model of star formation
, we use an average density profile that is a power law , @xmath65 , where @xmath66 is the distance from the center .
this is the density profile of the inner portion of a free falling isothermal sphere ( larson 1969 , 1973 , shu 1977 ) . using this density profile , the temperature of 100 k and the turbulent velocity dispersion equal to half of the free fall velocity
, we can calculate the imf and the star formation rate as a function of the distance from the center of the cloud during the collapse of this .
since the density is higher in the center than in the outer parts , the star formation rate is increasing towards the center .
this means that at some point during the collapse a bound stellar system could be formed around the center , while the remaining gas would be dispersed by supernova explosions .
the main aim of this work is to test if that bound system looks like a gc and if the scenario is consistent with observational constraints related to the formation of the galaxy .
the star formation process continues as long as the collapse is taking place , that is until the energy injected in the system by supernova explosions becomes larger than the gravitational energy of the cloud .
when this happens , if at all , a bound stellar system will be left only if the total mass of the stars is comparable or larger than the mass of the dispersed gas ( lada , margulis & dearborn 1984 ) . = 1.0 we compute the star formation rate and the imf during the collapse - i.e , at different densities . at any given density , we compare the total gravitational energy of the cloud with the kinetic energy injected into the gas motions by explosions of stars more massive than 8 m@xmath2 .
every supernova is supposed to liberate an energy of @xmath67 ergs that is turned into kinetic energy of the surrounding gas with an efficiency of 1% ( examples of efficiency estimates are given in spitzer 1978 , brown , burkert & truran 1995 ) . in fig .
1 we show that the total energy of the gaseous cloud becomes positive when the average cloud density is equal to 1.0 @xmath68 . at this density
the collapse is stopped and the gas is dispersed away from the stars , marking the end of the star formation process . in order to check if a bound stellar system is formed at all , we compute the star formation efficiency ( mass star / total mass ) at different distances from the center of the cloud ; the bound stellar system is formed wherever the efficiency is close to 100% , in fig .
2 the star formation efficiency is plotted versus a radial coordinate ( mass fraction ) .
the efficiency falls below 50% at a mass fraction of about 0.004 , that corresponds to a radius of 50 pc .
therefore _ our star formation model applied to a primordial cloud of @xmath69 m@xmath2 predicts the formation of a bound stellar system with a mass of @xmath70 m@xmath2 and a radius of 30 pc _ .
the integration of the function plotted in fig .
2 shows that approximately 0.002 of the total mass of the cloud has turned into a bound stellar system and 0.02 into stars located outside the bound system , that will evaporate and eventually form the halo of the galaxy ( without including the bulge component ) .
the imf and its typical mass are given by the formula 1 and 6 ( section 3 ) .
the imf for gcs is plotted in fig 3 ( dashed line ) .
the imf has a single maximum , an exponential cut off for smaller masses and a long intermittent tail for larger masses . in the figure we plot with the continuous line the exponent ( @xmath71 ) of the power law imf that approximates the predicted imf .
the slope for masses below 1 m@xmath2 is always smaller than the salpeter , @xmath72 , and the maximum is located at 0.1 m@xmath2 .
the slope becomes equal to the salpeter one around 2 m@xmath2 ( not shown in fig 3 ) .
therefore our model predicts that _ deep luminosity functions observed in gcs should be characterised by a cut off at luminosities corresponding to masses of about 0.1 m@xmath2 preceded by a progressive flattening of the distribution_.
the picture of gc formation proposed in the present work implies the following scenario for galaxy formation : * the protogalaxy is an ensemble of a few hundred primordial clouds of about @xmath0 m@xmath2 that are in self collapse .
* one gc is formed inside the nucleus of each cloud . * the remaining part of the cloud forms the halo stars that evaporate from the cloud . *
the cloud gas that is expelled out by the supernova explosions falls dissipatively into a disk system .
as we mentioned above , most of the stars formed outside the bound system are evaporated from the cloud .
the total mass in this stellar component is about 10 times larger than the mass of the gc .
moreover , the total mass of the dispersed gas is around 500 times the mass of the gc .
therefore our model predicts reasonable mass ratios between gcs and galaxy masses ( harris & racine 1979 ) and between gcs and the stars in the external halo ( woltjer 1975 , bahcall & soneira 1982 , harris & racine 1979 ) .
moreover , it also predicts the imf for halo stars and the initial metallicity of the galactic disk .
the halo imf is shown in fig 4 .
= 1.0 the most probable mass of halo stars is 6 times more massive than the one in gcs .
this is due to the fact that halo stars are formed at a lower density in more external regions of the primordial cloud compared to gc stars .
so we predict that _ deep stellar counts of halo stars will show a cut off in the imf below 0.6 m@xmath2_. the gas that falls into the disk system is previously metal - enriched by the star formation in the protoglobular clouds .
therefore we expect to find virtually _ no disk stars with metallicity lower than [ o / h]@xmath73_.
in this paper we have presented a model for the formation of the halo population of stars ( bulge excluded ) and gcs ( metallicity below 0.1 the solar value )
. the visible mass of such population in the galaxy is a few percent of the total galactic mass for the halo stars and a few per thousands for the gcs .
these mass ratios are reproduced by the present model for reasonable values of the non - baryonic mass of the universe .
searle & zinn ( 1978 ) and zinn ( 1985 ) have proved the existence of two distinct populations in the system of galactic gcs .
the two populations differ in metallicity and kinematic properties .
the halo gcs show no metallicity gradient outside 8 kpc from the galactic center while the disk gcs show a metallicity gradient and a minimum metallicity [ fe / h]=-1 .
several authors have proposed distinct mechanisms for the origin of the two gcs populations ( searle & zinn 1978 , zinn 1985 , rosenblatt , faber & blumenthal 1988 , van den bergh 1993 ) , while no attempt has been made here to model the formation of the disk gcs .
the halo gcs have been explained as the result of the star formation process in primordial mass fluctuations in a cdm spectrum ; halo stars are also formed in this way .
the typical mass of primordial protoglobular clouds is explained by assuming that the onset of h@xmath8 cooling is essential for the collapse and fragmentation of the baryonic gas in gravitational potentials dominated by collisionless dark matter .
in fact , it is the short cooling time in the gas that makes the shocks isothermal , so that purely baryonic strong density enhancement can be achieved , and the collapse and fragmentation of the baryonic gas can proceed .
= 1.0 we have found that the ratio of baryonic to non - baryonic mass , @xmath74 plays an important role in the formation of the combined system of halo gcs and halo stars . if @xmath75 too massive gcs are formed , in contradiction with the observations . on the other hand the formation of the estimated value of the mass of halo stars requires @xmath76 .
we can conclude that the primordial origin of halo gcs and halo stars in the cdm scenario is feasible only if @xmath77 , that is the density parameter @xmath78 , if the baryonic density parameter @xmath79 ( kolb & turner 1990 ) . our statistical model of star formation
is based on a fragmentation mechanism controlled by the statistics of fluid turbulence .
the universal character of the statistical properties of turbulence generate a functional form of the stellar imf which does not depend on environmental physical parameters , even though the typical stellar mass does .
the gcs imf has its low mass cutoff at 0.1 m@xmath2 .
the slope of the imf ( x=1 at 0.8 m@xmath2 ) is smaller than the salpeter one ( x=1.35 ) , and becomes comparable with it at about 2 m@xmath2 .
there have been attempts to determine the imf in gcs .
fahlman et al .
( 1989 ) , richer et al .
( 1991 ) and piotto & ortolani ( 1991 ) find very steep imfs .
nevertheless these determinations are affected by several uncertainties , the most important being the unknown mass - luminosity relation .
moreover the measured imfs span only a very limited range in masses ( 0.2 - 0.7 m@xmath2 ) in the best case , making any extrapolation for other masses very uncertain .
recent hst observations find a cutoff in the imf at about 0.1 m@xmath2 , in good agreement with our prediction .
models of gcs evaporation and core collapse ( hut & djorgovsky ( 1992 ) ) are consistent with observations if the average stellar mass is assumed to be 0.3 m@xmath2 , consistent with our model .
the expected non - baryonic dark matter inside the gc is only a small fraction of the gc mass because the high velocity ( about 100 km / s ) of the dark matter particles prevents them from falling into the baryonic central density condensation . as for the baryonic dark matter
, our imf contains a negligible number of brown dwarfs and jupiterian objects .
therefore the total dark matter content of gcs is expected to be small , in agreement with the observations .
the predicted stellar masses in the halo are 6 times larger than in the gcs .
this is due to the smaller average density of the gas where halo stars are formed ( the maximum in the halo imf corresponds to a density of 250 @xmath80 and in the gc of 10000 @xmath80 ) .
the halo stellar population is also more massive than the solar neighbourhood one , suggesting that the star formation process in the early protogalaxy is characterised by the presence of a large number of supernovae , compared with the subsequent star formation in the disk .
therefore the contribution to brown dwarfs and jupiters from the halo stars is also negligible .
the halo imf implies that in the evolution of every galaxy there will be a high blue luminosity during the first 2 gyr .
this event should be included in models of galactic luminosity evolution to be used for the interpretation of deep galaxy counts .
we predict a halo imf that is less steep than the salpeter s one below 2 m@xmath2 .
richer & fahlman ( 1992 ) observed low mass stars down to 0.14 m@xmath2 in the halo and suggested that the slope of the imf is instead steeper than the salpeter s one . on the other hand , more recent hst observations by bahcall et al .
1994 of high latitude fields found fewer faint red stars .
halo gcs in large galaxies show a dispersion in metallicity that is much larger than the internal one ( harris and racine ( 1979 ) ) ; this fact is sometimes used as an argument against the primordial origin of gcs , because it is believed that self - enrichment in the protoglobular cloud should lead to a heterogeneous metallicity distribution inside the gc . nevertheless in our picture the stars responsible for the enrichment
are formed during the dissipative collapse of the baryonic cloud , which takes a few @xmath0 years .
this time scale is long enough to allow for efficient mixing if only the central region of the cloud , where the gc will be formed , is considered .
the bulk of gc stars are formed in a period of only @xmath81 years ( because of the stellar feedback that destroys the cloud ) , when the central average density has become large enough , and therefore such stars can not be responsible for any further self - enrichment . on the other hand
the efficient mixing for the whole cloud requires a longer time than for the nucleus , so that some stars will be formed outside the gc with very low metallicity .
in fact the metallicity distribution of halo stars is known to extend to practically zero metallicity ( beers & sommer - larsen ( 1995 ) ) .
= 1.0 only a few supernovae are required to enrich the primordial cloud and achieve the value for the metallicity observed in halo gcs .
the small number of supernovae easily explains the range in metallicity observed in halo gcs , as due to statistical fluctuations in the number of supernovae events , before the final star formation burst .
our scenario for gc formation is not only consistent with metallicity determinations in gcs but also with the observed lower limit for the disk metallicity : there are virtually no disk stars found with metallicity lower than [ o / h]@xmath82 ( van den bergh 1962 , schmidt 1963 , beers & sommer - larsen 1995 ) .
we calculate the oxygen yield from the star formation in the protoglobular cloud , and find that the metallicity of the gas , that is dispersed from the cloud and falls into the disk , is about [ o / h]@xmath82 . so we do not expect any disk star to be formed with metallicity lower than this value , in agreement with the observations .
the observed abundances of oxygen over iron in halo stars ( matteucci & greggio 1986 , wheeler , sneden & truran 1989 , mcwilliam & rich 1994 ) , and the fact that the median halo stars metallicity ( norris 1986 ) is about the same as the median metallicity in gcs , suggest that halo stars and gcs belong to the same star formation process as stated in our model .
the main conclusions of this paper are the following : * gcs can be formed out of the nucleus of a primordial cloud of a few @xmath0 m@xmath2 of baryons . *
the mechanism of @xmath3 formation selects this typical mass scale out of a cdm spectrum . * most halo stars in the galaxy
( bulge excluded ) can be formed together with the ( halo ) gcs in the same clouds ; the halo and the gcs masses are sensitive to the ratio of baryonic over non baryonic matter , x@xmath83 , in the cloud . * a ratio @xmath84 produces realistic values for the mass of gcs ; if @xmath85 too large gcs are formed . * a ratio @xmath86 is needed in order to originate most of the halo stars in the protoglobular clouds during the formation of gcs . * the gas that is blown away from the disrupted protoglobular cloud by supernova explosions is used in the following star formation episodes ; therefore the initial disk metallicity ( minimum stellar metallicity in the disk ) is [ o / fe]@xmath73 .
* the imf in gcs has a cut off below 0.1 m@xmath2 while the halo imf has the cut off below 0.6 m@xmath2 .
* during the first 2 gyr of its life each galaxy has higher blue luminosity than normal disks .
this is due to the fact that star formation ( halo stars ) in the protoglobular cloud produces a significant number of massive stars , which also means an epoch of large number of supernova explosions .
we have shown that this scenario is consistent with several observations .
the star formation rate and the stellar imf are essential ingredients of the model . in order to predict them
we have used a new physical model of star formation . in this model the functional form of the imf is shown to be universal because it is based on the statistics of turbulence which are universal properties of flows .
future work should concentrate in the attempt of simulating the collapse of cdm primordial protoglobular clouds .
cosmological simulations on such small scales are at the moment beyond the capabilities of available computer resources , since they require a very large dynamical range .
we thank s.m . fall and a. kashlinsky for fruitful comments and discussions during the elaboration of this manuscript .
we also enjoyed discussions with c. flynn , c. lacey & b. pagel .
we are thankful to the referee , dr .
joe silk , for helpful comments on the paper .
99 aguilar , l. , hut , p. , ostriker , j.p .
1988 , apj , 335 , 720 arp , h.c . ,
baum , w.a . ,
sandage , a.r .
1952 , aj , 57 , 4 ashman , k. e. 1990 , mnras , 247 , 662 ashman , k. e. , zepf , s. e. 1992 , apj , 384 , 50 bahcall , j.n . , soneira , r. 1984 , apjs , 277 , 27 bahcall , j.n . , flynn , c. , gould , a. , kirhakos , s. 1994 , apj , 435 , l51 beers , t. c. , sommer - larsen , j. 1995 , apjs , 96 , 175 brown , j.h . , burkert , a. , truran , j.w .
1991 , apj , 376 , 115 brown , j.h . , burkert , a. , truran , j.w .
1995 , apj , 440 , 666 carlberg , r.g . 1981 ,
mnras , 197 , 1021 charlton , j. c. , laguna , p. 1995 , apj , 444 , 193 duerr , r. , imhoff , c.l . , lada , c.j .
1982 , apj , 261 , 135 dove , j.e . , mandy , m.e .
1986 , 311 , l93 fahlman , g.g .
, richer , h.b . ,
searle , l. , thompson , i.b .
1989 , apj , 343 , l49 fall , s. m. 1978 , globular clusters fall , s. m. , rees j. m. 1985 , apj , 298 , 18 gunn , j. e. 1977 , apj , 218 , 592 hanes , d. e. 1977 , _ mem .
_ ras , 84 , 45 harris , w. e. 1991 , ara&a , 29 , 543 harris , w. e. 1995 , apj harris , w. e. , racine , r. 1979 , ara&a , 17 , 241 harris , w. e. , pudritz , r. e. 1994 , apj , 429 , 177 hirasawa , t. 1969 , prog .
phys . , 42 , 523 hirasawa , t. , aiza , k. , taketani , m. 1969 , prog .
phys . , 41 , 835 hut , p. , djorgovski , s. 1992 , nature , 359 , 806 hutchins , j.b . 1976 ,
apj , 205 , 103 iben , i. , renzini , a. , 1984 , phys . rep .
105 , 329 jacquin , l. , cambon , c. , blin , e. 1993 , phys .
fluids a 5 , 2539 jimenez , r. , thejll , p. , jrgensen , u. , macdonald , j. , pagel , b. 1996 , mnras , _ in press _ kang , h. , shapiro , p.r . , fall , s. m. , rees , m. j. 1990 , apj , 363 , 488 kolb , e.w . ,
turner , m.s .
1990 , the early universe , addison - wesley kumai , y. , basu , b. , fujimoto , m. 1993 , apj , 404 , 144 lada , c. j. , lada , e. a. , clemens , d. p. , bally , j. 1994 , apj , 429 , 694 larson , r. b. 1969 , mnras , 145 , 271 larson , r. b. 1969 , mnras , 145 , 405 larson , r.b . 1973 , fund .
cosm . physc .
1,1 larson , r.b .
1986 , in nearly normal galaxies , ed .
s. faber ( ny : springer ) p.26 larson , r. b. 1990 , in physical processes in fragmentation and star formation , ed . r. capuzzo - dolcetta , c. chiosi , a. di fazio , p.389 .
dordrecht : kluwer lepp , s. , shull , j. m. 1983 , apj , 270 , 578 matteucci , f. , greggio , l. 1986 , a&a , 154 , 279 matsuda , t. , sato , h. , takeda , h. 1969 , prog theor .
phys . , 42 , 219 miller , g. e. , scalo , j. m. 1979 , apjs , 41 , 413 mooney , t.j . , solomon , p.m. 1988 , apj , 334 , l51 myers , p.c . , dame , t.m . ,
thaddeus , p. , cohen , r.s . ,
silverberg , l.f . ,
nordlund , .
, padoan , p. 1996 , submitted to physics of fluids .
norris , j. 1986 , apjs , 61 , 667 dwek , e. , hauser , m.g .
1986 , apj , 301 , 398 mcwilliam , a. , rich , r.m .
1994 , apjs , 91 , 749 ostriker , j. p. 1986 , in the harlow - shapley symposium on globular clusters systems in galaxies , eds .
grindlay , j.e . and
davis philip , a.g .
palla , f. , zinnecker , h. 1988 , in iau symposium 126 : the harlow - shapley symposium on globular clusters systems in galaxies , ed .
grindlay and a.g.d .
philip ( dordrecht : reidel ) , p. 323 palla , f. , salpeter , e.e . , stahler , s.w .
1983 , apj , 271 , 632 padoan , p. 1995 , mnras , 277 , 377 padoan , p. , nordlund , .
p. , jones , b. j. t. 1996 , submitted to apj .
padoan , p. , jones , b. j. t. , nordlund , .
, submitted to apj .
nordlund , .
, padoan , p. 1996 , submitted to phys .
fluids peebles , p.j.e . , dicke , r.h .
1968 , apj , 154 , 891 peebles , p.j.e . , 1984 , science , 224 , 1385 richer , h.b .
, fahlman g.g . , buonanno , r. , fusi pecci , f. , searle , l. , thompson , i.b .
1991 , apj , 381 , 147 richer , h.b .
, fahlman g.g .
1992 , nature , 358 , 383 rosenblatt , e. i. , faber , s. m. , blumenthal , g. r. 1988 , apj , 330 , 191 rotman , d. 1991 , phys .
fluids , a 3 ( 7 ) , 1792 schmidt , m. 1963 , apj , 137 , 758 searle , l. , zinn , r. 1978 , apj , 225 , 357 shapiro , p. , kang h. 1987 , 318 , 32 shu , f.h .
1977 , apj , 214 , 488 silk , j. 1977 , apj , 211 , 638 spitzer , l. jr .
1978 , in physical processes in the interstellar medium ( new york : john wiley ) vandenberg , d.a .
, bolte , m. , stetson , p.b .
1990 , aj , 100 , 445 vietri , m. , pesce , e. 1995 , apj , 442 , 618 van den bergh , s. 1962 , aj , 67 , 486 van den bergh , s. 1993 , apj , 411 , 178 wheeler , j.c . ,
sneden , c. , truran , j.w . , 1989 , ara&a , 27 , 279 woltjer , l. 1975 , a&a , 42 , 109 zinnecker , h. , keable , c.j . , dunlop , j.s . ,
cannon , r.d .
, griffiths , w.k .
in globular clusters systems in galaxies , iau symp .
j. grindlay , a.g.d .
philip , p. 603 .
dordrecht : reidel | using a new physical model for star formation ( padoan 1995 ) we have tested the possibility that globular clusters ( gcs ) are formed from primordial mass fluctuations , whose mass scale ( @xmath0 - @xmath1 m@xmath2 ) is selected out of a cdm spectrum by the mechanism of non - equilibrium formation of @xmath3 .
we show that such clouds are able to convert about 0.003 of their total mass into a bound system ( gc ) and about 0.02 into halo stars .
the metal enriched gas is dispersed away from the gc by supernova explosions and forms the galactic disk .
these mass ratios between gcs , halo and disk depend on the predicted imf which is a consequence of the universal statistics of fluid turbulence .
they also depend on the ratio of baryonic over non - baryonic mass , @xmath4 , and are comparable with the values observed in typical spiral galaxies for @xmath5 .
the computed mass and radius for a gc ( @xmath6 m@xmath2 and 30 pc ) are in good agreement with the average values in the galaxy .
the model predicts an exponential cut off in the stellar imf below 0.1 m@xmath2 in gcs and 0.6 m@xmath2 in the halo .
the quite massive star formation in primordial clouds leads to a large number of supernovae and to a high blue luminosity during the first two gyr of the life of every galaxy .
epsf star : formation globular clusters : formation galaxy : formation |
we first address garcia _ et al . _
s concerns about our online survey @xcite , which they suggest induced a positivity bias in respondents answers .
garcia _ et al .
_ claim that a set of function words in the liwc ( language inquiry and word count ) data set @xcite show a wide spectrum of average happiness with positive skew ( their fig 1a ) when , according to their interpretation , these words should exhibit a dirac delta function located at neutral ( @xmath0=5 on a 1 to 9 scale ) .
we expose and address two fundamental errors .
first , function words in the liwc data set are simply not emotionally neutral .
the liwc data set annotates 4487 words and stems on a wide range of dimensions @xcite .
we find a total of 421 words and 48 stems are coded as function words with 450 matches in our data set when using stems . of these ,
only 7 are indicated as emotional ( 5 positive , 2 negative ) which appears to support garcia _ et al .
_ s interpretation .
however , a straightforward reading of the liwc list of function words reveals that these words readily bear emotional weight as exemplified by `` greatest '' and `` worst '' .
we present some of the most extremely and most neutrally rated liwc function words in tab .
[ tab : mhl - reply.liwc - function ] .
more generally , `` not looking at the words '' and `` not showing the words '' are pervasive issues with word- and phrase - based summary statistics for texts .
we should be able to see how specific words contribute to summary statistics for texts to provide ( 1 ) an assurance the measure is performing as intended , and ( 2 ) insight into the text itself .
for example , all sentiment scoring algorithms based on words and phrases must be able to plainly show why one text is more positive through changes in word frequency , such as through the word shifts we have developed for both print @xcite and as interactive , online visualizations @xcite .
elsewhere , in studying the google books corpus , we have produced analogous word shifts for the jensen - shannon divergence @xcite .
we exhort other researchers to produce similar word shifts ( and not just word clouds ) , and to question work with no such counterpart .
second , as we discuss in detail in sec .
[ sec : mhl - liwc.freqdepend ] below , no statement about biases can be made about sets of words chosen without frequency of usage incorporated .
any given set of words may have a positive , neutral , or negative bias , but we must know how they are chosen before being able to generalize ( as we have done thoroughly in @xcite ) . because we have no guarantee that the expert - generated liwc function words are exhaustive and because we are merging words of highly variable usage frequency , a finding of an average positive bias for liwc function words is meaningless , regardless of their transparent capacity for being non - neutral . .
three subsets of 450 liwc function words with high , neutral , and low average happiness scores from our labmt study @xcite ( stems provide more matches than those found by garcia et al . ) .
each word s score is the average rating for 50 participants ( scale is 1 to 9 with 1 = most negative , 5 = neutral , and 9 = most positive ) .
function words may carry emotional weight and can not be presumed to be neutral . [ cols=">,>",options="header " , ] emotional words in liwc provide another case in point . around 20% of the liwc data set ( 907 words and stems )
are denoted as having positive affect ( 160 words and 247 stems ) or negative affect ( 151 words and 349 stems ) .
while stems complicate word counts , the liwc data set clearly does not show evidence of a positivity bias .
because the liwc data set is expert - curated and meant to be general , it does not fit any natural corpora with respect to usage frequency ( i.e. , liwc words constitute an unsystematic sampling ) .
word lists meant to accurately reflect statistical properties of language must be built directly from the most frequently used words of well defined corpora a point we will return to several times in this reply .
an earlier expert - curated word list , the smaller anew data set @xcite , similarly fails in these respects , showing a fairly flat distribution across average happiness @xcite .
liwc , along with all word data sets , should not be considered an unimpeachable `` gold standard '' ; language is far too complex to make such an assured statement .
all word data sets , including our own , will have limitations .
we next contend with a comparison made by garcia _
et al . _ between our work on english with a similar sized survey by warriner and kuperman ( wk ) @xcite .
wk generated a merged list of 13,915 english words , the bulk of which ( 11,826 ) are a list of lemmas taken from movie subtitles .
immediately , we have a mismatch : our word list incorporated the 5000 most frequently used words ( or tokens ) in each of four disparate corpora ( new york times , google books , music lyrics , and twitter ) whereas wk s list is mostly lemmas ( e.g. , `` sing '' but not `` sung '' or `` sang '' ) taken from one coherent corpus .
further , each word was scored by 50 participants in our study , compared with 1420 for the wk study . in their fig . 1b , garcia _ et al .
_ show histograms for the two word lists , which seem to indicate more negative words in the wk list and a higher median word happiness for our word list .
but such a comparison is unsound : the words behind each histogram are not the same and word frequency is not being controlled for .
the two histograms can not be sensibly compared , and we can discard garcia _
s finding that the median level of average word happiness @xmath0 for our full data set is 0.28 above the median level for the wk data set . nevertheless , garcia _ et al . _
do appropriately compare the shared subset of words found in both data sets , finding a much smaller difference between median values of @xmath0 of 0.07 .
they then suggest that our use of cartoon faces to indicate the 1 to 9 scale of happiness responses induces a positive bias in respondents choices , referencing a study that found a non - smiling face to be slightly negative @xcite .
their claim lacks foundation for several reasons .
first , wk employed a reverse 9 point scale , with 1 = happy and 9 = unhappy , flipping the scores after completing the surveys ( also used in @xcite ) .
we have no objection to wk s approach but evidently this further complicates any comparison between the two studies .
indeed , one might reasonably hypothesize that flipping the direction of the ratings could be the sole cause of the minor discrepancy between the words scored by both studies .
second , we gave all participants clear written instructions that 5 was neutral . in wanting to generate results that could be compared with existing work
, we followed the design of bradley and lang in their anew study @xcite , who used both cartoon figures in their self - assessment manikins and written ( spoken for anew ) instructions ; we departed only in orienting positive to the right .
( as have many others , garcia _ et al .
_ have used the anew study in their research @xcite . ) wk take pains to compare their scores with the anew study ( which they use in part as a control ) and other studies , finding their results are `` roughly equivalent '' ( p. 6 ) . and as we noted in @xcite , basic function words which are expected to be neutral such as `` the '' and `` of '' were appropriately scored as such , indicating that the survey mechanism we used was not adding a simple positive shift .
third , given the nature of language and surveys and changing demographics online , an exact match for the medians would be a remarkable achievement . the agreement between the labmt ( our english word set @xcite ) and wk is still a strong one , and we show scatter plots for the matching word happiness scores in fig .
[ fig : mhl - reply.wk - comp ] for labmt , anew , and wk .
visually , we see the three studies are sympathetic with each other , particularly when we acknowledge the typical standard deviation for the scores of individual words ( on the order of 1 to 2 ) .
we used reduced ( or standard ) major axis regression @xcite to obtain the fits shown in fig .
[ fig : mhl - reply.wk - comp ] , @xmath1 .
we see that anew s scores grow slightly faster than that of both labmt and wk ( @xmath2 = 1.08 and 1.07 ) and wk similarly relates to labmt ( @xmath3 ) .
fourth , and rather finally , according to the argument of garcia _ et al .
_ regarding faces , the median for anew should be higher than that of wk ( noting again that they used the same happy to sad directionality ) , yet we see the _ opposite _ ( 5.29 versus 5.44 ) . moreover
, comparing medians alone is insufficient our regression analysis shows that , for the words they have in common , wk appears more emotionally biased than labmt with @xmath2 = 1.04 .
we note that a much richer comparison could be carried out at the level of individual ratings , but this is far too detailed for the present response .
we turn now to garcia _ et al .
_ s central claim : that we claimed to find that a positivity bias is _ independent _ of word frequency across 10 languages .
in fact , we instead variously stated that a positivity bias is `` strongly '' and `` largely '' independent of frequency , and we explored the minor departures from pure independence in detail for all 24 corpora across 10 languages ( see @xcite and the paper s online appendices ) .
garcia _ et al .
_ write that our paper specifically conflicts with two previous works , their own @xcite and that of warriner and kuperman @xcite .
we are able to dismiss @xcite due to it being founded on a misapplication of an information theoretic formula by piantatosi et al .
@xcite , and which we demonstrate elsewhere @xcite . setting aside this misrepresentation , garcia _ et al .
_ s issue with our work becomes to what degree frequency independence is followed , and they provide an alternative analysis of how positivity behaves with usage frequency . whereas we performed the regression @xmath4 where @xmath5 is rank
, they claim @xmath6 is more appropriate . once again
, we stand by our own principled analysis for the following reasons .
* 1 . mismatch of scored word list and word list with frequency : * in attempting to say anything about a given quality of words as it relates to usage frequency within a specific corpora , a complete census of words by frequency must be on hand .
_ have taken our merged word lists for each language and applied them to data sets for which they do not necessarily fit .
problematically , their word lists do not contain ranks , and consequently there are words missing in uncontrolled ways from the data they perform regression on . for the example of english , our 10,222 words will ( likely ) match the most common words in any sufficiently large english corpus .
but the matching becomes more peculiar the rarer the word , and the inclusion of twitter in our word list means garcia _ et al .
_ have found `` lolz '' , `` bieber '' and `` tweeps '' in google books . in fig . [
fig : mhl - reply.jellyfish]a , we plot average happiness as a function of frequency of usage for the word list they created from google books .
the scatter plot is clearly unsuitable for linear regression .
we show an estimate of cumulative coverage at the bottom ( see caption ) , which crashes soon after reaching 5000 words . *
rank is an appropriate variable to regress happiness ( or any word quality ) against : * garcia _ et al . _
state that regression against frequency @xmath7 is a better choice because information is lost in moving to rank @xmath5 .
however , the general adherence of natural language to zipf s law , @xmath8 , provides an immediate counterargument @xcite , even acknowledging the possibility of a scaling break @xcite .
because word usage frequency is so variable , great care must be taken with any analysis . as we show for the case of english google books in fig .
[ fig : mhl - reply.jellyfish]a , regression on @xmath9 will be gravely compromised by the increasing preponderance of words at lower frequencies ( a common issue with measuring power - law slopes ) , and , based on even the words for which coverage is reasonable , it would evidently be in poor judgment to extrapolate from any linear fit across frequencies .
by contrast , fig .
[ fig : mhl - reply.jellyfish]b shows how usage rank is perfectly suited for regression , and is the basis for the `` jellyfish '' plots we provided in fig . 3 of @xcite and in the paper s online appendices .
our jellyfish plots make the general conformity to a rough ( we do not claim `` physical - law '' strict ) scale independence abundantly clear . by using rank
, we are able to perform a much finer analysis than garcia et al .
propose , and we show in all corpora that the deciles for a sliding window of 375 ranks changes at most rather slowly . finally , in fig .
[ fig : mhl - reply.jellyfish]c , we present how @xmath10 behaves as a function of @xmath11 , illustrating both the error of choosing @xmath9 and that our results will be essentially unchanged if we regress against @xmath11 . in closing , we emphasize that minor deviations from frequency independence remain a secondary aspect of our observation that the polyanna hypothesis holds for a diverse set of languages , and are wholly irrelevant for the instrumental aspect of our work in creating text - based hedonometric tools .
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 @noop * * , ( ) , @noop ( ) , @noop ( ) , @noop `` , '' ( ) @noop ( ) , @noop * * , ( ) @noop * * , ( ) , @noop @noop `` , '' ( ) , @noop _ _ , ( , , ) @noop * * ( ) @noop ( ) , @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) , @noop _ _ , patterns ( , , ) | we demonstrate that the concerns expressed by garcia _
et al .
_ are misplaced , due to ( 1 ) a misreading of our findings in @xcite ; ( 2 ) a widespread failure to examine and present words in support of asserted summary quantities based on word usage frequencies ; and ( 3 ) a range of misconceptions about word usage frequency , word rank , and expert - constructed word lists . in particular , we show that the english component of our study compares well statistically with two related surveys , that no survey design influence is apparent , and that estimates of measurement error do not explain the positivity biases reported in our work and that of others .
we further demonstrate that for the frequency dependence of positivity of which we explored the nuances in great detail in @xcite garcia _ et al .
_ did not perform a reanalysis of our data they instead carried out an analysis of a different , statistically improper data set and introduced a nonlinearity before performing linear regression .
note : the present manuscript is an elaboration of our short reply letter @xcite . |
through the last several years , quantized function algebras have attracted a lot of attention ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and many others ) . amongst these , @xmath2 has attracted the most attention . since in fact
a number of candidates for the quantized function algebra of @xmath7 matrices have been proposed , we stress that the one we consider here is the `` original '' ( or `` standard '' or `` official '' ) one introduced by faddeev , reshetikhin , and takhtajan in @xcite .
we wish to consider some natural subalgebras and quotients of this algebra , namely the subalgebra @xmath8 of quantized @xmath0 matrices , the subalgebra @xmath9 obtained by removing the @xmath10 corner generated by those @xmath11 for which @xmath12 , and finally the quotients @xmath13 obtained by factoring out the ideal @xmath14 generated by all @xmath5 quantum subdeterminants
. the emphasis will throughout be on the case where @xmath15 is a primitive root of unity .
the major tool is the theory developed by de concini and procesi in @xcite as well as the theory of p.i
. algebras .
we have found it convenient to collect these results , some corollaries to them , as well as some further developments in section [ pi ] following immediately after this introduction . in some sense
, the results of de concini and procesi turn the problem into an elementary one , but which at the same time is of a such kind that one should not expect general results except possibly in special cases .
indeed , a major part of the procedure is to bring into block diagonal form an integer coefficient skewsymmetric form .
the new results we present relate to ( iterated ) skew polynomial extensions and are particularly useful for the algebras @xmath9 .
even for the known case ( c.f .
@xcite ) @xmath2 they are sufficient , and we have chosen to illustrate this in section [ stand ] . actually , the case of @xmath2 was brought to a completion by the discovery of a very special phenomenon for the associated quasipolynomial algebra ( @xcite ) and a substantial further development of this observation now makes it possible to attack @xmath8 .
this is done in section [ nr ] .
section [ qm ] is devoted to the proof of the isomorphism @xmath16\simeq m^{r+1}_q(n)[d^{-1}]$ ] , where @xmath17 is the quantum @xmath18 minor corresponding to the subalgebra generated by those @xmath11 for which @xmath19 .
a major tool is the representation theory of quantized enveloping algebras .
having established that , we turn our attention to @xmath9 in section [ qfa ] .
indeed , for questions relating to degree , center , etc . , it is sufficient to consider this algebra , which is more manageable .
the methods that worked well for @xmath8 do not apply as easily but , fortunately , the results obtained in section [ pi ] are applicable , especially after some fortunate guesses relating to the center .
as one consequence we obtain ( combine corollary [ com - cor ] with theeorem [ com - thm ] ) : _ if @xmath15 is an odd primitive @xmath6th root of unity then @xmath20_. we thank ken goodearl and tom lenagan for criticizing the original proof of proposition [ in - pro ] .
in this section , unless explicitly stated , @xmath21 denotes a prime p.i .
algebra and @xmath22 an algebraically closed field @xmath22 of characteristic @xmath23 .
we assume throughout that @xmath21 is finitely generated ( affine ) as an algebra over @xmath22 .
we start by recalling some basic results from the theory of p.i .
algebras and then we show these results can be applied to calculations of the degree of an algebra .
let us first recall some basic definitions .
an algebra @xmath21 is said to be of degree at most @xmath17 , if @xmath21 satisfies all identities of @xmath24-matrices over a commutative ring .
if no such @xmath17 exists @xmath21 is said to have infinite degree . in either case
the smallest possible @xmath17 is denoted by @xmath25 .
degree of an algebra @xmath21 , @xmath26 , is @xmath27 , where @xmath28 is the smallest possible degree of a multilinear polynomial which vanishes on @xmath21 . if @xmath21 is a prime p.i .- algebra , such as ours , one gets @xmath25 = @xmath26 . for an affine prime p.i .
algebra @xmath21 one has several useful results concerning the ( p.i.)-degree . to state and prove these , let us first recall that the intersection of all primitive ideals of @xmath21 is @xmath23 ( * ? ? ?
* theorem 13.10.3 ) and all primitive ideals are maximal by kaplansky s theorem ( * ? ? ?
* theorem 13.3.8 ) .
@xmath29 , where @xmath30 runs through the set of all maximal two - sided ideals . _
proof : _ since @xmath31 is a factor algebra of @xmath21 for all maximal ideals @xmath30 , we get that the right hand side is bounded by the left hand side . from (
* corollary 13.6.7 ) we get that @xmath32 if and only if @xmath33 , the standard identity , is an identity for @xmath31 , thus @xmath34 implies that @xmath33 is an identity for @xmath31 for all @xmath30 .
therefore @xmath33 is an identity for @xmath35 , where @xmath30 ranges over all maximal ideals .
but by the above remarks , @xmath21 has a natural embedding into @xmath36 and hence any identity of @xmath36 is also an identity of @xmath21 .
let @xmath30 be a maximal two - sided ideal of @xmath21 .
then @xmath37 is a simple p.i .
algebra and hence of the form @xmath38 , where @xmath39 is a division ring which is finite dimensional over its center @xmath40 .
moreover , @xmath41 ( @xcite ) .
if @xmath42 is a maximal commutative subfield of @xmath39 then @xmath43 since by ( * ? ? ?
* theorem 13.10.3 ( the proof ) ) it is finite dimensional over @xmath22 and the latter is algebraically closed . hence ( ( * ? ? ? * lemma 13.3.4 ) ) , @xmath44 for some @xmath45 .
thus we conclude [ piprop ] @xmath46 , where @xmath47 runs through all irreducible @xmath21-modules .
[ rem - deg ] the goldie quotient ring @xmath48 of @xmath21 can be obtained by inverting the non - zero _ central _ elements of @xmath21 ( * ? ? ?
* corollary 13.6.7 ) .
thus , @xmath21 and @xmath48 have the same p.i .
therefore , any ring @xmath36 between @xmath21 and @xmath48 has the same p.i .
degree as @xmath21 and in case @xmath36 is affine over @xmath22 , @xmath21 and @xmath36 have the same degree . from ( * ? ? ?
* corollary 13.3.5 ) we now get [ 5 ] @xmath49 , where @xmath50 denotes the quotient field of the center of @xmath21 .
as noted in @xcite we have [ pro - reg ] let @xmath51 be a finite set of regular elements of @xmath21 .
there exists an irreducible representation @xmath52 of @xmath21 of maximal degree in which @xmath53 are units . _
proof : _ let @xmath54 $ ] . then @xmath55 . by remark [ rem - deg ] the p.i
. degrees of @xmath21 and @xmath36 are equal .
let @xmath52 be an irreducible representation of @xmath36 of maximal dimension on a finite - dimensional vector space @xmath56 over @xmath22 .
let @xmath57 denote the restriction of @xmath52 to @xmath21 .
then @xmath58 for all @xmath59 .
moreover , each @xmath60 is a regular linear map , hence by the cayley - hamilton theorem its inverse is in @xmath61 .
thus @xmath62 and hence @xmath57 is irreducible .
a different approach to finding the degree of certain algebras has been found by de concini and procesi ( * ? ? ?
we recall some of their results : let @xmath63 be a skewsymmetric @xmath7 matrix such that @xmath64 .
given @xmath65 , the quasipolynomial algebra @xmath66 $ ] is the algebra over the field @xmath22 generated by @xmath67 and with defining relations @xmath68 we call @xmath65 the * defining matrix * of the quasipolynomial algebra . in the following , @xmath69
is always assumed to be a primitive @xmath6th root of unity .
de concini and procesi proved [ deppro ] @xmath70=\sqrt h$ ] , where @xmath71 is the cardinality of the image of the map induced by @xmath65 @xmath72 defined by @xmath73 , where @xmath74 denotes taking residue class in each coordinate .
furthermore @xmath66 $ ] is a free module over its center of rank @xmath75 . in (
* proposition , p.61 ) it was shown that @xmath76 $ ] is an azumaya algebra under the assumption that @xmath22 is an algebraically closed field of characteristic @xmath23 and arbitrary @xmath65 .
since the algebras @xmath77 $ ] all are affine prime p.i .
algebras we can use some of the results from above to prove that the assumptions on @xmath22 made by de concini and procesi are superfluous .
[ azu ] let @xmath22 be a field and @xmath65 a skewsymmetric matrix with integer coefficients .
the algebra @xmath76 $ ] is an azumaya algebra .
_ proof : _ we wish to use the artin - procesi theorem ( * ? ? ?
* theorem 13.7.14 ) .
thus we have to show that @xmath78 is an identity for @xmath21 if and only if @xmath78 is an identity for @xmath79 for all primes of @xmath21 ( ( * ? ? ?
* corollary 13.6.7 ) ) .
@xmath78 is an identity for @xmath21 or @xmath79 if @xmath78 vanishes on all @xmath6-tuples of monomials in the @xmath80s , @xmath81 .
for any such an @xmath6-tuple @xmath82 we get @xmath83 with @xmath84 depending on @xmath85 .
since no @xmath86 can belong to a prime @xmath87 , the claim follows .
many of the algebras considered in the following are iterated ore extension .
we therefore recall some results on the degree of iterated ore extensions or skew polynomial algebras . in ( * ?
* theorem 1 ) and in ( * ? ? ?
* , theorem ) it was proved that if @xmath88 is an affine prime algebra over a field of characteristic @xmath23 , then @xmath89=\deg r[\theta;\alpha]$ ] provided @xmath89 $ ] is finite . here
, @xmath90 is @xmath22-automorphism of @xmath88 and @xmath91 an @xmath90-derivation on @xmath88 . combining this result with proposition [ 5 ] , the degrees of the so called dipper - donkin algebras @xmath92 and the quantized ( `` official '' ) matrix algebras @xmath2 were found by jakobsen and zhang @xcite , ( for the quantum parameter @xmath15 a primitive @xmath6th root of unity ) . for later purposes we need a few more results concerning skew polynomial algebras ( implicitly in @xcite ) .
we consider an affine prime p.i .
algebra @xmath88 and a skew polynomial algebra @xmath93,\ ] ] where @xmath90 is a @xmath22-automorphisms of @xmath88 and @xmath91 an @xmath90-derivation .
we assume @xmath21 is a p.i .
algebra and @xmath90 has finite order .
notice that for any regular element @xmath94 ( resp .
@xmath95 ) @xmath96 $ ] ( resp
$ ] ) is a @xmath22-affine prime p.i .
subalgebra of the goldie quotient ring of @xmath88 ( resp .
@xmath21 ) hence by the previous results has the same degree as @xmath88 ( resp .
@xmath21 ) .
the next lemma follows easily , and is very well known .
@xmath98.\ ] ] combining these results for @xmath99 $ ] we obtain [ cid ] in case there exist regular @xmath100 in @xmath88 and @xmath101 in @xmath88 such that @xmath102 commutes with all elements of @xmath21 then @xmath103=\deg r.\ ] ] in ( * ? ? ?
* section 4 ) it was proved that such @xmath104 and @xmath105 exist when @xmath90 induces the identity on @xmath106 , the center of @xmath88 . in the present article this result
will only be used in situations in which there exist a regular element @xmath94 and an element @xmath107 such that @xmath108 is central in @xmath21 .
observe that @xmath4 is regular in @xmath21 . in an irreducible representation @xmath52 of @xmath21 of maximal degree , we may by proposition [ pro - reg ] assume that @xmath4 is invertible . then @xmath109 and
hence @xmath52 remains irreducible when restricted to @xmath88 .
thus , @xmath110 .
it was also shown in ( * ? ? ?
* the proof of theorem 3.1 ) that in case @xmath90 is not the identity on @xmath106 , then there exists a finitely generated multiplicatively closed @xmath90-invariant set @xmath111 of central elements of @xmath88 such that @xmath112[\theta;\alpha,\delta]\simeq r[t^{-1}][\theta';\alpha],\ ] ] where @xmath113 for some @xmath114)$ ] , and such that @xmath115 $ ] is @xmath22-affine . in case
there exists a subalgebra @xmath116 of the center @xmath117 of @xmath88 such that i ) @xmath117 is a finite @xmath116 module , ii ) @xmath118 , and iii ) @xmath119 , de concini and procesi proved ( ( * ? ? ?
* theorem p. 58 ) ) that @xmath120=(\deg r)\cdot k,\ ] ] where @xmath22 is the order of @xmath90 s restriction to @xmath106 .
[ by the methods of @xcite and @xcite one can in fact show that the special assumptions on @xmath121 are superfluous .
one just needs that @xmath122 $ ] is a prime p.i . algebra . ] in particular , we get , provided @xmath122 $ ] is a p.i .
algebra , [ coa ] let @xmath88 be a prime p.i .
algebra and @xmath90 an automorphism of @xmath88 of order @xmath22 .
if there exists an element @xmath123 such that the @xmath90 orbit of @xmath124 has order @xmath22 , then @xmath125)=(\deg r)\cdot k.\ ] ]
the `` standard '' quantized function algebra @xmath2 of @xmath7 matrices is the quadratic algebra generated by @xmath126 elements @xmath127 and with defining relations @xmath128 for @xmath129 . it is well known that the monomials @xmath130 for @xmath131 form a pbw - type basis of @xmath2 for any @xmath132 . the quantum determinant @xmath133is central for any @xmath134 . viewing @xmath2 as an iterated skew polynomial algebra ( c.f .
below ) , it follows that _ the associated quasipolynomial algebra _
@xmath135 is given in terms of the same generators , but with defining relations @xmath136 for @xmath129 .
later on , we shall encounter a number of subalgebras @xmath36 of @xmath2 . for each of these , analogously to the above , the associated quasipolynomial algebra @xmath137 is the algebra with the same generators but where in the defining relations , all terms of the form @xmath138 have been dropped .
the degree of @xmath2 was found in @xcite to be @xmath139 for @xmath15 an @xmath6th root of unity , @xmath6 odd .
the approach there utilized the result of procesi and de concini ( ( * ? ? ?
* theorem p. 59 ) ) according to which , as a special case , @xmath140
. we will reprove this result by utilizing the results of section [ pi ] . before doing so
let us introduce some notation , which will be used also later in this paper .
we view @xmath141 as the @xmath22-algebra generated by the elements @xmath11 , @xmath142 and we will then view @xmath2 as an iterated ore extension of @xmath141 obtained by adjoining the indeterminates as follows : @xmath143\cdots[z_{n , n-1 } ; \alpha_{n , n-1}\delta_{n , n-1 } ] [ z_{1,n};\alpha_{1n}]\cdots[z_{n , n } ; \alpha_{n , n},\delta_n],\ ] ] where @xmath144 for @xmath145 or @xmath146 , and where @xmath147 for @xmath145 or @xmath148 .
changing slightly the notation from parshall and wang ( * ? ? ?
* section 4 ) we define @xmath149 and @xmath150 and let @xmath151 ( the quantum determinant based on the rows in @xmath152 and columns in @xmath65 ) and @xmath153 .
later on , we shall introduce some more general elements with these names but letting @xmath154 in figure [ f1 ] , the elements @xmath155 in that figure are precisely what have been defined here ( with @xmath156 and @xmath157 ) .
in @xcite the following notion was introduced [ covdef ] an element @xmath158 is called covariant if for any @xmath11 there exists an integer @xmath159 such that @xmath160 clearly , @xmath161 and @xmath162 are covariant .
it was then shown that the elements @xmath163 and @xmath164 are covariant . utilizing this ,
the following elements were also found to belong to the center : @xmath165 for @xmath166 . for the convenience of the reader we list the covariance properties of the elements @xmath167 ( from which the centrality of @xmath168 and @xmath169 also is obvious ) .
let @xmath149 and @xmath150 as previously : @xmath170 it is well - known that @xmath2 ( being an iterated ore extension ) is a domain , thus the @xmath100 in lemma [ cid ] are automatically regular if non - zero .
[ detsec ] [ detdeg ] @xmath171 , where @xmath6 is an odd integer and @xmath15 is a primitive @xmath6th root of unity . _
proof : _ we use induction on @xmath45 . since the formula clearly holds for @xmath172 , it suffices to prove @xmath173 first notice that @xmath174 where in fact @xmath4 up to a sign is @xmath175 for @xmath141 , and where @xmath101 , when expanded in the pbw basis , does not contain @xmath176 either . by lemma [ cid ]
we see that @xmath177 is the same as the degree of the algebra where @xmath176 is excluded .
the same argument works for @xmath178 .
one just has to replace @xmath175 by @xmath179 in the procedure .
thus , @xmath180\cdots[z_{n , n-1};\alpha_{n , n-1},\delta_{n , n-1}].\ ] ] let @xmath181 be the algebra obtained by adjoining @xmath182 to @xmath141 .
let @xmath183 and @xmath184 be the quantities in @xmath141 analogous to @xmath185 and @xmath186 , then notice that @xmath187 is a central element in @xmath181 and @xmath188 therefore by proposition [ coa ] and because @xmath6 is odd , @xmath189
. in the case of @xmath190 one may just use @xmath191 as @xmath192 .
since @xmath193 , there is a similar conclusion . the proof is thus completed .
we consider here the quadratic algebra @xmath8 consisting of @xmath0 matrices ( @xmath194 ) .
we determine the degree and the center in case @xmath6 is `` good '' and also obtain insight , in some cases even full , into the cases where @xmath6 is not so `` good '' .
[ [ section ] ] we assume that @xmath195 and @xmath196 with @xmath197 odd and @xmath101 as big as possible .
we display @xmath101 central elements of @xmath8 .
( for a hint of how these were discovered , see the proof of proposition [ bl2 ] below
. also recall from @xcite that the centres of @xmath8 and @xmath198 are in bijective correspondence via the leading term . )
we begin by defining elements @xmath199 for @xmath200 .
first set @xmath201 .
@xmath202 is the quantum @xmath203 minor involving the rows @xmath204 and columns @xmath205 ( for @xmath206 ) .
then consider @xmath207 and let @xmath208 be the quantum @xmath18 determinant involving the rows @xmath209 .
finally , @xmath210 is the quantum @xmath211 determinant involving columns @xmath212 and rows @xmath213 for @xmath214 . for @xmath215 the elements @xmath216 defined by @xmath217 are central .
_ proof : _ let us consider the case @xmath218 .
we may then view our @xmath0 matrix as being built up of @xmath219 blocks @xmath220 of size @xmath101 , block @xmath221 consisting of rows and columns @xmath222 , block @xmath223 consisting of rows @xmath224 and columns @xmath222 , etc .
let us now look at some @xmath225 .
due to the covariance of the various determinants it is possible to see that the commutativity of @xmath226 with @xmath227 is equivalent to picking up a factor of @xmath228 for each each block @xmath229 and each @xmath230 and that indeed the whole computation may be viewed as the computation for commutativity of the analogous expression computed in @xmath231 .
here it is a matter of investigating the matrix @xmath232 given by @xmath233 and checking that @xmath234 is central .
but since @xmath235 and @xmath236 are odd , this is straightforward .
indeed , the computation is reduced to ascertaining that if @xmath237 is odd and @xmath238 then @xmath239 in @xmath240 .
the case @xmath241 is similar though slightly more complicated .
one may view the relevant diagram over the minors as being obtained by a translation along the line from the upper right corner to the lower left . only near the boundaries
does one need to check carefully that the signs match up .
the details are left to the reader , c.f .
figure [ fig6 ] . finally observe that one has the following result , the first part of which is just as in @xcite : [ 5.222 ] i ) let @xmath6 be even and @xmath15 a primitive @xmath6th root of unity .
any element of the form @xmath242 where @xmath243 and @xmath244 , is in the center of @xmath198 .
\ii ) if @xmath245 is even then @xmath246 the analogous elements of @xmath8 ( i.e. the same expressions but where the generators are in that algebra ) are also central in this case
. it follows by an argument similar to the one in ( * ? ? ?
* theorem 6.2 ) that in case @xmath6 is `` good '' ( c.f .
( [ goodlabel ] ) and proposition [ gcd ] below ) then the elements @xmath247 together with the elements @xmath248 generate the center .
[ [ section-1 ] ] following @xcite , the defining matrix @xmath65 of the associated quasipolynomial algebra @xmath198 , with respect to a natural basis @xmath249 , is in fact the matrix of the map @xmath250 where @xmath251 for @xmath252 .
we start by computing the rank of this map or , equivalently , the dimension of the kernel .
let @xmath253 thus , if @xmath6 is `` good '' , e.g. a large prime , @xmath254 indeed , @xmath255 by ( * ? ? ?
* theorem p. 59 ) , and the latter is given by theorem [ deppro ] .
the cardinality of @xmath65 is clearly the same as for any block diagonal form of @xmath65 .
now consider a @xmath256 matrix @xmath257 for some @xmath258 .
considered as a map from @xmath259 to @xmath260 , the image clearly has cardinality @xmath261 , where @xmath262 denotes the greatest common divisor of @xmath6 and @xmath4 .
this means that if e.g. @xmath6 is a prime bigger than all non - zero elements in a block form of @xmath65 then each non - trivial block contributes with a factor @xmath6 to the total degree . but clearly , there are @xmath263 such blocks .
recall from @xcite that @xmath264 where @xmath265 .
in particular , @xmath266 .
now observe that @xmath267 where @xmath268 is invertible , indeed , @xmath269 .
this observation will also be used later , but it follows immediately that the kernel is given by those @xmath0 matrices @xmath21 for which @xmath270 if @xmath271 a straightforward computation gives that ( [ rela ] ) is equivalent to @xmath272[ext ] if we define , for all @xmath273 , for @xmath274 , and for @xmath275 @xmath276 then ( [ rela ] ) is equivalent to @xmath277 [ gcd ] let @xmath278 .
specifically , let @xmath279 and @xmath280 .
then @xmath65 is non - invertible if and only both @xmath235 and @xmath236 are odd . in this case , @xmath281 _ proof : _ observe that by definition , @xmath235 and @xmath236 can not both be even .
now , according to ( [ newform2 ] ) and ( [ trel ] ) @xmath282 thus , if a solution to ( [ rela ] ) is to exist with @xmath283 then @xmath284 . in this case , a solution is given by @xmath285 and all other @xmath286 .
more generally , a non - zero solution exists if and only if @xmath284 . in this case
there are @xmath101 independent solutions given by @xmath287 the claim now follows .
we wish to obtain more precise information about the blocks of a diagonal form of the associated matrix of the quasipolynomial algebra @xmath198 .
[ bl1 ] in the case where @xmath288 there is an irreducible module of dimension @xmath289 where @xmath290}.\ ] ] _ proof : _ in this case the term @xmath291 disappears and so @xmath292 .
hence , by covariance , in an irreducible module , @xmath11 is either zero or invertible . consider the subalgebra @xmath293 by proposition [ pro - reg ] there is an irreducible representation of maximal dimension of this algebra in which all the generators are invertible .
given such an irreducible module , the recipe @xmath294 where @xmath295 , @xmath296 , and the @xmath297s are arbitrary constants , define an irreducible representation of @xmath198 .
the proof is thus completed if we can establish that there are precisely @xmath298 $ ] non - trivial blocks in the block diagonal form of the associated matrix of @xmath299 ( and hence that the degree is @xmath300}$ ] ) . for this purpose ,
let @xmath301 for @xmath302 and let @xmath303 for @xmath304 . upon the replacements @xmath305 ( @xmath306 ) , @xmath307 , and @xmath308 ( @xmath309 ) , the pair @xmath310 decouples completely leaving us with an algebra which is isomorphic to @xmath311 .
it is well - known ( and elementary ) to see that there are @xmath312 $ ] non - trivial blocks in the block diagonal form corresponding to @xmath313 .
the result follows directly from these observations .
actually , we did not use anything about the algebra except that it was contained in a box of size @xmath0 , hence we get the following corollary to the proof : let @xmath314 be a subalgebra of @xmath8 such that @xmath315 and such that @xmath316 . then , in case @xmath288 there is an irreducible module of dimension @xmath289 where @xmath290}.\ ] ] for general @xmath15 we get a result similar to proposition [ bl1 ] . specifically , given a representation of @xmath317 in which the generators are denoted @xmath318 and in which @xmath319 is invertible , the recipe @xmath320 defines a representation of @xmath2 as can be seen by a straightforward but tedious computation .
[ [ section-2 ] ] we wish to take a closer look at the degree and center in the case where @xmath6 is not necessarily a prime and where @xmath321 are arbitrary . for this purpose
we need more information about the pertinent block diagonal form .
[ bl2 ] the non - trivial blocks in a block diagonal form of the defining matrix @xmath65 of @xmath198 are either of the form @xmath322 , @xmath323 , or of the form @xmath324 .
_ proof : _ we will begin by studying the center of @xmath198 at a primitive @xmath6th root of unity . if @xmath21 is an @xmath0 integer matrix , the condition for a monomial @xmath325 to be in the center is precisely ( in the notation of ( [ 44 ] ) ) that @xmath326 returning to ( [ m - eq ] ) , it follows that @xmath327 where @xmath328 is an integer matrix whose entries all are multiples of @xmath6 .
however , basically due to the @xmath329 in @xmath330 , not all such matrices @xmath30 need define a solution @xmath21 to ( [ mod - eq ] ) .
returning now to the equations ( [ newform1 ] - [ trel ] ) , these remain valid when reinterpreted as equations _
modulo @xmath331_. in case @xmath284 , with @xmath332 as in proposition [ gcd ] , we just get the old solutions possibly with some elements of the form ( [ 521 ] ) or ( [ 522 ] ) superimposed .
but the case @xmath333 now is possible since we just need @xmath334 .
thus , it should be proportional to @xmath335 . in all cases
it follows that the entries of @xmath21 are integer multiples of @xmath335 and hence that the element @xmath336 satisfies that its fourth power is in the central subalgebra generated by the elements @xmath337 . but
suppose that there is a block of the form @xmath338 with @xmath339 .
then there are monomials @xmath340 such that @xmath341 and such that @xmath336 commutes with all other generators .
but then the element @xmath342 is central for any @xmath6 which is a multiple of @xmath101 and this is a contradiction since @xmath343 will not be in the above mentioned central subalgebra .
[ remmm ] it follows from ( [ symeq ] ) that if @xmath336 is central , then so is @xmath344 for any @xmath345 with @xmath346 .
this symmetry can be used to construct new solutions from given ones , c.f .
below . [ 4.6 ] the non - trivial blocks in a block diagonal form of the defining matrix @xmath65 of @xmath198 are : @xmath347 matrices of the form @xmath322 and @xmath348 matrices of the form @xmath323 or @xmath324 . _
proof : _ this follows immediately from proposition [ azu ] , proposition [ bl1 ] , and proposition [ bl2 ] .
we finish with some remarks about the occurrence of `` @xmath349 ' 's : suppose that @xmath350 are relatively prime .
following the reasoning just before remark [ remmm ] , if we are to have a genuine solution to ( [ symeq ] ) involving @xmath351 then by ( [ trel ] ) we must have @xmath352 @xmath353 @xmath354 for all @xmath355 .
we may assume that @xmath352 @xmath356 for all @xmath357 and all @xmath358 .
this follows since if some @xmath352 @xmath359 then , by lemma [ 5.222 ] the @xmath360is part of a central element involving @xmath361 , and @xmath362 and may thus be discarded .
we furthermore assume that @xmath363 for @xmath364 and @xmath365 for @xmath366 , where each @xmath367and @xmath368 is @xmath23 or @xmath369 modulo @xmath349 .
consider first a pair of indices @xmath370with @xmath371 .
then the condition for @xmath372to commute with @xmath373 in the quasipolynomial algebra is @xmath374 by subtracting consecutive terms it follows that @xmath375 also observe that @xmath376 and @xmath377 at a point @xmath378 with @xmath379 we get , utilizing the parity properties , @xmath380 likewise , at @xmath381 with @xmath382 we get @xmath383 and , finally , at @xmath384we get @xmath385 for these equations to have solutions , @xmath4 and @xmath45 must have the the same parity
. if both are odd , there are no further restrictions for these equations to have a solution which also solves ( [ symeq ] ) .
returning to the general situation , if both @xmath321 are even , it turns out to be a further necessary ( and also sufficient ) condition for ( [ symeq ] ) to have a genuine @xmath335 solution that they are equal modulo @xmath349 .
it should now be observed that if there is a general center , that is , if @xmath65 is singular , amongst these solutions there will be some which are actually of no interest , namely those solutions that correspond to a zero on the right hand side of ( [ symeq ] ) .
let us now assume that @xmath4 is prime .
if @xmath386 , @xmath45 odd does not contribute by the above analysis .
if it is even , it is forced to be of the form @xmath387 which means that the central elements already have been picked up by the general central elements .
suppose then that @xmath4 is an odd prime .
if @xmath4 does not divide @xmath45 , is it also forced to be odd and hence @xmath65 is singular and hence the solutions we pick up are general central elements raised to the power @xmath335 .
this takes care of all cases except @xmath388 for some positive integer @xmath389 .
if @xmath389 is odd we have an @xmath4-dimensional center : again nothing new .
finally , if @xmath389 is even , the previous elements do not give anything .
however , there are in fact some non - trivial @xmath351-central elements .
specifically , let @xmath390 be the matrix whose non - zero coefficients @xmath391 satisfy @xmath392 and add to that the matrix @xmath393 whose whose non - zero coefficients @xmath394 satisfy @xmath395 then @xmath396 defines a central element .
moreover , using the symmetries of the original equation , we get in fact @xmath397 solutions ( which clearly is as much as could be hoped for ) .
we thus have the following partial result let @xmath4 be a prime .
then @xmath324 occurs in the block diagonal form of the defining matrix @xmath65 of @xmath198 if and only if @xmath4 is odd and @xmath398 for some even integer @xmath389 . in this case
, there are @xmath399 such blocks .
for each @xmath400 , @xmath401 denotes the ideal generated by all @xmath402 quantum minors .
we consider here the function algebra of rank @xmath4 matrices .
specifically , let @xmath403 for each @xmath404 , let @xmath405 denote the @xmath406 quantum determinant of the subalgebra generated by the elements @xmath11 with @xmath407 .
set @xmath408 .
the natural candidate for quantized rank @xmath4 matrices is then @xmath409,\ ] ] where we shall return to the issue of inverting @xmath17 shortly .
we wish to compare this algebra to a somewhat more manageable one , namely @xmath9 , where the algebra @xmath9 is the subalgebra of @xmath2 generated by those @xmath11 for which @xmath410 .
in @xcite , goodearl and lenagan proved that @xmath14 is completely prime .
moreover , rigal proved that @xmath411\simeq m^2_q(n)[d^{-1}]$ ] ( @xcite ) .
we shall prove below that @xmath412\simeq m^{r+1}_q(n)[d^{-1}]\ ] ] for a general @xmath4 and on the way give a new proof of the former result . for ( [ sense ] ) to make sense
we first of all need the following : [ 5.2 ] @xmath408 is regular in @xmath413 both in the case of @xmath15 generic and the case where @xmath414 is a primitive @xmath6th root of unity .
_ proof : _ it is proved in ( * ? ? ?
* theorem 2.5 ) that @xmath413 is a domain .
since @xmath17 clearly is non - zero in @xmath413 we get the result .
we now offer an alternative argument in the generic case .
we need this result to prove ( [ sj - star ] ) in case @xmath15 is generic .
later on , we also obtain ( [ sj - star ] ) in the case of a primitive @xmath6th root of unity by ring theoretic methods .
_ proof of proposition [ 5.2 ] for @xmath15 generic : _ our proof relies on representation theory .
first of all , for this case it was proved in @xcite that @xmath2 is a bi - module of a version @xmath415 of the quantized enveloping algebra of @xmath416 .
essentially , this version is what results if one starts from the quantized serre relations and view the @xmath15 entering there as a complex number .
furthermore , it is assumed that @xmath417 and that @xmath15 is not a root of unity .
the results obtained by @xcite reveal that the same general picture holds as in the well - known case for @xmath418 @xcite . specifically , each @xmath419 is a @xmath415 sub - bi - module .
moreover , there is a decomposition @xmath420 as a bi - module . here
, each @xmath421 is an irreducible @xmath422 module .
the highest weight vector in @xmath423 is given by @xmath424 for @xmath425 and @xmath426 .
for each @xmath427 let @xmath428 .
then the weight @xmath429 of the @xmath430 in ( [ v - l ] ) is given by @xmath431 .
there are no multiplicities .
furthermore , each @xmath432 is invariant . if @xmath433 denotes the direct sum of the highest weight modules whose highest weight vectors are of the form @xmath434 with @xmath435 , then this is precisely equal to @xmath436 . we are now ready to prove that @xmath17 is regular : @xmath437 .
assume that @xmath438 .
with no loss of generality we may assume that @xmath439 and @xmath440 .
observe that @xmath441 is a primitive vector for all @xmath132 .
the 2-sided action of the borel sualgebra @xmath442 at @xmath15 generic then preserves the general form of ( [ de - co ] ) and hence we may assume that @xmath443 is a sum of highest weight vectors of different highest weights , and using weight considerations , we may assume that @xmath443 is a single highest weight vector .
this is then of the form ( [ v - l ] ) with @xmath444 .
but we must still have that @xmath445 and this is a contradiction .
below we show that the powers of @xmath17 can be inverted in a manageable manner .
let @xmath446 be the natural homomorphism from @xmath9 to @xmath447 .
[ orepro ] let @xmath448 .
then @xmath449 is an ore set of regular elements in @xmath1 . _
proof : _ since @xmath450 is regular , it suffices to prove that @xmath314 is an ore set in @xmath2 . clearly , @xmath314 is an ore set in @xmath451 since @xmath17 is central in that algebra .
the remaining indeterminates are now added in a suitable order , i.e. in such a way that @xmath2 is an iterated ore extension of @xmath452 . if @xmath453 denotes the automorphism corresponding to @xmath11 , then either @xmath454 or @xmath455 .
the result then follows by ( * ? ? ?
* lemma 1.4 ) . in case @xmath15 is a primitive @xmath6th root of unity @xmath456 is central by ( * ? ? ?
* lemma 7.23 ) and the localization is then a central localization .
it follows now that @xmath446 induces a homomorphism @xmath457:a_{n , r}[s^{-1}]\mapsto m_q^{r+1}[s^{-1}]$ ] , which is onto since for each @xmath458 or @xmath459 there exists a @xmath460 such that @xmath461 is a quantum @xmath462 minor
. we can now prove [ in - pro]the natural homomorphism @xmath463 is injective : _ proof : _ we first give the details for the `` @xmath15 generic '' case . under the bi - module action of @xmath415 on @xmath2 ,
the algebra @xmath9 is invariant under a borel subalgebra from one side and under the opposite borel subalgebra from the other side . since @xmath14 is invariant
we may argue exactly as in the proof of proposition [ 5.2 ] .
thus , if there is a non - zero element @xmath464 such that @xmath465 , there is also a non - zero highest weight vector @xmath466 such that @xmath467 .
but then @xmath468 is of the form ( [ v - l ] ) with @xmath469 and by looking at leading terms in these quantized determinants , this is easily seen to be impossible .
next suppose @xmath15 is a primitive @xmath6th root of unity . here
all the algebras @xmath470\,,\;m_q^{r+1}(n ) \mbox { and } m_q^{r+1}(n)(d^{-1})\ ] ] are prime affine p.i . algebras .
we recall that for such algebras krull - dimension , transcendence degree and g - k - dimension coincide ( * ? ? ?
* proposition 10.6 ) .
in the sequel , @xmath471 denotes one of these . by the definition of transcendence degree any affine algebra between the original algebra , @xmath36 , and
the simple quotient algebra will have the same dimension as @xmath36 . by (
* proposition 6.5.4 ) @xmath472\le n^2-(n - r)^2\,,\ ] ] because it is a localization of a @xmath473 iterated ore extension . by a recent result @xcite @xmath474 and therefore @xmath475\ge n^2-(n - r)^2\,.\ ] ]
if @xmath476\ne 0 $ ] then the krull - dimension of @xmath457(a_{n , r}[s^{-1}])$ ] must be strictly less than that of @xmath477 $ ] , therefore @xmath457 $ ] is injective and so is @xmath446 .
[ co1]for @xmath15 either generic or a primitive root of unity , @xmath478\simeq m^{r+1}_q(n)[d^{-1}].\ ] ] [ com - cor ] if @xmath15 is a primitive root of unity , @xmath479
in this section we determine the degree of @xmath9 .
the general strategy of the proof is similar to the proof of theorem [ detdeg ] .
we begin by adjoining the variables @xmath486 to @xmath487 .
let @xmath488 denote the algebra obtained by adjoining @xmath489 so that @xmath490 .
we show that there exist suitable central elements @xmath491 in @xmath487 which behave nicely under the automorphisms induced by each of the variables @xmath486 .
this makes it possible to construct a central element for each @xmath488 which has an @xmath6th order orbit under @xmath492 .
thus , an application of proposition [ coa ] is possible with the conclusion that the degree of @xmath493\cdots[z_{r , n};\alpha_{r , n},\delta_{r , n}]$ ] is @xmath494 times the degree of @xmath487 .
after that we construct @xmath4 central elements of @xmath9 . in the general situation
they will have the same shape as those for @xmath487 .
for each @xmath495 there will in fact be a central element of @xmath9 which is an inhomogeneous polynomial of degree 1 in that variable .
thus , lemma [ cid ] implies that the degree does not go up by adjoining @xmath496 to @xmath497 .
having established this , the theorem is proved .
we will first consider the cases where @xmath498 .
extend the previous definitions of @xmath499 as follows : for @xmath500 , let @xmath501 denote the quantum determinant of the sub - algebra generated by @xmath502 elements @xmath506 are defined analogously for @xmath507 by interchanging all @xmath11 with @xmath508 in @xmath504 .
it follows easily that the @xmath186 and @xmath185 are covariant for all @xmath509 .
notice that @xmath510 in case @xmath11 belongs to the subalgebra involved in defining @xmath501 .
also , if @xmath11 is not in the subalgebra used to define @xmath501 then @xmath511 if @xmath512 and @xmath513 .
similarly , @xmath514 in case there exists a @xmath515 used in the definition of @xmath501 such that @xmath516 or @xmath517 for some positive integer @xmath235 . in order to make the arguments and ideas as clear as possible , we will first consider the case where we , starting from @xmath518 , determine the degree of @xmath519 .
we begin by adjoining the elements @xmath520 to @xmath521 in order of increasing first index .
thus , we get a sequence of skewpolynomial algebras @xmath522 for @xmath523 . for convenience ,
let @xmath524 .
now , since @xmath525 we can use those central elements of the latter that do not involve @xmath526 . in this way we get
@xmath4 central elements : @xmath527 where we write @xmath528 for @xmath529 . since @xmath534 is central in @xmath535 , and
since the length of the orbit of @xmath536 s action on @xmath534 is @xmath6 , we get that adjoining @xmath537 raises the degree by a factor @xmath6 .
next , @xmath538 is clearly central in @xmath539 and @xmath540 .
thus , since @xmath6 is odd , we get that when we adjoin @xmath541 to the previously constructed algebra the degree again goes up by a factor @xmath6 .
first of all , there are @xmath548 central elements coming from @xmath549 : @xmath550 the remaining central element @xmath534 has got to involve @xmath551 . we claim that the element @xmath552 fulfills the requirements .
first of all , clearly @xmath553 for all @xmath554 . to prove commutativity in all details would involve checking that the five factors of @xmath534 are in such a balance with each other that the @xmath555 , while being the product of five terms of the form @xmath556 and while each @xmath556 depends on the actual form of @xmath557 , in the end always equals @xmath558 .
we leave the somewhat tedious ( and somewhat amusing ) details of this , as well as similar claims later on , to the reader .
we can now start adjoining the elements @xmath559 .
since for each @xmath560 , @xmath559 occurs in the summands of @xmath561 to either the power 1 or 0 , lemma [ cid ] applies and the degree remains unchanged . in the cases
@xmath568 there are no central elements coming from @xmath2 , but there are 2 types of central elements still yielding a total of @xmath4 central elements . specifically for each @xmath569 with @xmath570
we have the central elements @xmath571 the remaining central elements where @xmath572 can be gotten by the following : @xmath573 now let us finally comment on the case of a general @xmath574 : here one can easily construct @xmath4 central elements from the previous recipes
. indeed , observe that each previously constructed central element contains exactly one pair of factors @xmath575 which are not full @xmath18 quantum determinants .
if , say , @xmath576 is @xmath577 then @xmath578 is @xmath579 .
even more precisely , @xmath580 and @xmath581 .
then a number of factors of the form @xmath582 and @xmath583 are inserted for @xmath584 - as long as the resulting indices are positive
. clearly this procedure continues to work for a general @xmath45 .
we omit the finer details and refer to figure [ fig3 ] .
that concludes the proof .
the non - trivial blocks in a block diagonal form of the defining matrix @xmath585 of @xmath586 are : @xmath587 matrices of the form @xmath322 and @xmath588 matrices of the form @xmath323 .
let @xmath589 if @xmath6 is odd and @xmath590 if @xmath6 is even .
then , in particular , the degree is @xmath591 . | first some old as well as new results about p.i .
algebras , ore extensions , and degrees are presented
. then quantized @xmath0 matrices as well as certain quantized factor algebras @xmath1 of @xmath2 are analyzed . for @xmath3 , @xmath1 is the quantized function algebra of rank @xmath4 matrices obtained by working modulo the ideal generated by all @xmath5 quantum subdeterminants and a certain localization of this algebra is proved to be isomorphic to a more manageable one . in all cases ,
the quantum parameter is a primitive @xmath6th roots of unity .
the degrees and centers of the algebras are determined when @xmath6 is a prime and the general structure is obtained for arbitrary @xmath6 . [ section ] [ thm]definition [ thm]remark [ thm]lemma [ thm]corollary [ thm]proposition [ section ] |
thin films of metal nanoclusters with a large fraction of atoms exposed to the surface are promising candidate from technological point of view .
their thermodynamic , optical , electronic , magnetic and mechanical properties are quite different from their bulk counterparts @xcite .
one of the most efficient ways to exhibit these properties @xcite is the preparation of monodispersed metal nanoclusters on a specific substrate . for this purpose
size - selected nanoclusters with a low or zero deposition energy are ideal @xcite as was felt over the years .
isolated islands of nano dimension on surfaces is of immense importance in recent days @xcite . due to quantum confinement effect
these dots exhibit many properties ( mostly superior ) which are far away from their bulk counterparts .
these nanodimensional dots having sizes below their bohr diameters are known as quantum dots which are worth studying in every material from the fundamental point of view .
formation of isolated islands or dots of specific sizes is important not only from fundamental aspect but also for many technological possibilities regarding memory , size and speed of future devices .
supported metal clusters of specific sizes are important to study for their catalytic properties @xcite .
ionized cluster beam deposition ( icbd ) is one of the methods to get nanosized dots on a substrate in a controlled way . to get
continuous deposition of size selected clusters with controlled impact energy and with appreciable cluster flux magnetron based gas aggregation types nanocluster source introduced by haberland et al.@xciteequipped with qmf is superior to other sources @xcite .
haberland et al .
@xcite showed deposition of clusters in different energy regime yields different film morphologies . by molecular dynamic ( md ) simulation
they showed that low energy deposition keeps shapes and sizes of preformed clusters intact after deposition .
such low energy or near to zero energy deposition is known as soft - landing . in tune with md simulations , we report here the formation of isolated nano - scale dots of the sizes comparable to the mass - selected clusters in flight by soft landing of ionized copper clusters on amorphous carbon films .
a uhv ( ultra high vacuum ) deposition system ( oxford applied research , uk ) with gas aggregation type magnetron based nanocluster source has been used for production of copper nanoclusters .
after the clusters are formed in the aggregation region , they are passed through a quadrupole mass filter that is capable of analysing clusters to a desired size with a narrow window of mass spectrum . for the present study ,
amorphous carbon films coated on a standard 300 mesh copper grid ( ni grid for energy dispersive x - ray spectroscopy study ) were used as substrates for deposition .
argon and helium gases were used for magnetron sputtering , aggregation and carrying the clusters to deposition chamber from the nanocluster source as no extraction arrangement is there in the system[described elsewhere ] .
base pressure before deposition was @xmath1 mbar and during deposition pressure in deposition chamber and in magnetron chamber were@xmath2 mbar and @xmath3 mbar respectively and cluster ion current was maintained at @xmath4 na controlling different parameters which controls cluster production .
substrates were kept at zero volt ( grounded ) to avoid any acceleration of the positively charged clusters before deposition .
substrate surface plane was normal to the incident cluster beam direction for all depositions .
clusters of diameter @xmath0 nm ( mass @xmath5 amu i.e. @xmath6 atoms , using wigner - seitz radius for copper , @xmath7 = 0.147 nm , detailed calculations in @xcite ) were selected for deposition with qmf resolution below @xmath8 .
care was taken to eliminate sources of kinetic energy of the clusters as much as possible .
all the deposited samples on the same day of deposition were investigated using transmission electron microscope ( tem ) facility ( fei tecnai g2 s - twin ) operating at 200 kv . for the information about composition of deposited film ,
the energy dispersive x - ray analysis ( edx ) of the films was performed in the same tem using 200 kv electron beam .
figure 1 shows the morphological properties of the deposited films .
the tem images presented in fig .
1(a - d ) are taken for different time of deposition under same conditions ( i.e , same ion current measured at qmf , magnetron power , parametersof mass filter etc . ) .
the corresponding histograms of the tem images are shown below them in fig 1 h(a - d ) .
each histogram shows a very narrow size distribution around 3 to 4 nm with an increase with time of deposition or with surface coverage . at lowest coverage or for 20 min of deposition
, distribution is centred at @xmath0 nm of diameter which is the selected diameter of the clusters before deposition .
figure 2(a ) shows a tem image from the 20 min deposited sample .
a close observation shows that almost all the tem features are circular i.e , the clusters are spherical in shape . for longer time of deposition
( e.g. about 1 hour ) , ramified islands are formed as shown in fig .
some of the features are indicated here to show that these features of tem image of the 60 min deposited sample are not simple circular , instead they are ramified . for compositional information of films ,
the energy dispersive x - ray ( edx ) analysis was performed using the attached edx facility with tem .
2(c ) shows the edx spectrum , depicting strong peaks of c , ni and cu .
island density vs deposition time plot in fig .
2(d ) shows that initially island density increases linearly but it becomes saturated after certain value of coverage ( here coverage is defined as percentage of surface area covered by clusters out of total surface area ) .
this plot has been obtained calculating the island density for all the deposited sample in different time .
island density means number of island per unit area and when this obtained number is multiplied by the cluster projected area which supposed to be circular of diameter equal to the diameter the cluster then it becomes the number of island density per cluster site .
this term certainly increases with time of deposition but if there is some saturation in the plot it gives the signature of surface diffusion and aggregation .
the value of saturation island density measures the diffusion .
if the value is very high then diffusion is not so high but there is a signature of diffusion as the island density is being saturated . in the present study island density
was found to saturate at coverage below @xmath9 and it is very low@xcite than ideal substrate condition where the substrate has no defect sites and also the saturation island density value is quite high .
moreover the ramified shapes are not well branched or big in size . in the case of surface diffusion defect sites on the surface
traps the clusters and their movement ceases @xcite yielding to very low diffusion coefficient .
following bardotti _
@xcite using the saturation island density in the equation@xmath10 , ( where @xmath11 is surface diffusion coefficient , @xmath12 is saturation island density per cluster site , @xmath13 = 0.336 @xcite , @xmath14 is incident cluster flux and @xmath15 is cluster diameter ) , the surface diffusion coefficient in room temperature comes out to be equal to @xmath16 , which is very low compared to other experiments where graphite substrate were used .
the in - set of 2(d ) is derived from the average value of coverages at different time plotted against the time of deposition .
this plot shows a linear monotonous increase in coverage with of time .
this confirms the fact that deposited films are of submonolayer thickness and clusters are not falling on the previously deposited clusters or islands .
the linear increase also gives the clear indication of the experimental fact that a constant flux has been maintained through all deposition experiments .
positively charged clusters which were deposited on the substrate @xmath17 cm away from the magnetron target were free from the negative bias applied to the target for sputtering and so as mentioned earlier that there is no other force to change the kinetic energy of the clusters except gas flow and so clusters attained very low kinetic energy near the soft landing .
soft - landed clusters are not pinned to the substrate surface and so they diffuse on the substrate . in deposition of cluster via soft - landing process the initial stage of growth of a film can be best described by the _ dda model _
@xcite , as deposition of incoming clusters takes place via diffusion and aggregation .
@xcite showed that a critical size of coalescence determines that the island will be formed by total coalescence or pure juxtaposition .
critical size of coalescence is that size of the clusters below which total coalescence occurs and clusters bigger than that do not coalesce but stick together to form ramified islands . in conclusion
, isolated islands have been formed with controllable size by soft - landing cu - nanoclusters on amorphous carbon film .
the phenomenon is explained by different models of cluster deposition . in the present case for large copper clusters ( 1200 atoms / cluster ) , the diffusion is very slow as compared to other experiments @xcite . as a result
monodispersed film formation has been possible with nearly soft landing deposition as exhibited in the present study .
detailed investigations with other cluster materials and substrates in this direction are in progress .
haberland h , in : springer series in chemical physics 52 ( clusters of atoms and molecules i , theory , experiment and clusters of atoms ) ( 1994 ) chapter 1 and 3 p. 1 and 207
binns c , surf sci rep 2001 ; 44 : 1 - 49 .
bhattacharyya s r , datta d , shyjumon i , smirnov b m , chini t k , ghose d et al , j. phys .
2009 ; 42 : 035306(1 - 9 ) .
palmer r e , pratontep s , boyen h -g , nature mater .
2003 ; 2 : 443 - 8 .
kashtanov p v , hippler r , smirnov b m , bhattacharyya s r , europhys .
2010 ; 90:16001(p1-p4 ) .
harbich w , in meiwes - broer k h ( ed . ) , _ metal clusters at surfaces _ , springer , berlin , 2000 .
heiz u , bullock e l , j mater chem 2004 ; 14 : 564 - 77 .
jensen p , rev .
phys 1999 ; 71 : 1695 - 735 .
tainoff d , bardotti l , tournus f , guiraud g , boisron o , and mlinon p , j. phys .
c 2008 ; 112 : 6842 - 49 .
h. haberland h , mall m , mosseler m , qiang y , rainers t , turner y , j vac sci technol a 1994 ; 12 : 2925 - 31 .
majumdar a , kpp d , ganeva m , datta d , bhattacharyya s r , hippler r , rev .
2009 ; 80 : 095103(1 - 6 ) .
haberland h , insepov z , moseler m , phys .
b 1995 ; 51 : 11061 - 7 .
francis gm , goldby i m , kuipers l , issendorff b von , palmer r e , j. chem .
soc 1996 ; dalton trans : 665 - 71 .
carroll s j , seeger k , palmer r e , appl .
lett 1997 ; 72 : 305 - 7 .
bardotti l , jensen p , hoareau a , treilleux m , cabaud b , phys .
lett 1995 ; 74 : 4694 - 7 .
jensen p , barabsi a -l , larralde h , havlin s , stanley h e , phys .
b 1994 ; 50 : 15316 - 29 .
yoon b , akulin v m , cahuzac ph , carlier f , de frutos m , masson a , mory c , colliex c , brchignac c , surf .
sci 1999 ; 443 : 76 - 88 .
* figure captions : * + * fig.1 . :
* ( upper panel ) shows the transmission electron microscopy ( tem ) images of the deposited cu - nanocluster films on carbon - coated ni grids at different times of deposition . the deposited times are 20 , 40 , 50 and 60 min presented in ( a ) , ( b ) , ( c ) and ( d ) respectively .
the histograms of the images are shown in the lower panel as h(a ) to h(d ) corresponding to the images ( a ) to ( d ) respectively . + *
: * ( a ) the typical tem image for 20 min deposition showing circular spot from spherical clusters , while for longer time ( 60 min ) deposition ramified structures along with elongated features are seen in ( b ) . in ( c )
energy dispersive x - ray ( edx ) spectrum shows the composition of the film on tem grid for 60 min deposition sample .
the island density of each deposition time is presented in fig .
the in - set in this figure shows the coverage ( % ) of the substrates by cu - clusters as a function of deposition time . | deposition of size - selected metal nanoclusters on a substrate with very low kinetic energy helps to keep the clusters intact with respect to their shape and size as compared to clusters in flight condition . here
we report formation of isolated monodispersed islands of copper of desired size on carbon film by deposition of size selected copper clusters ( @xmath0 nm ) in soft - landing method .
copper clusters were produced by a magnetron based gas aggregation type source equipped with a quardrupole mass filter ( qmf ) to select size of clusters before landing .
transmission electron microscopy ( tem ) study shows that diffusion of islands is very low . , nanoclusters , soft - landing , copper clusters , surface diffusion of islands , icbd .
pacs : 68.37.lp , 68.35.md , 68.55.a- |
silica ( sio@xmath1 ) is one of the most important and widely used materials in today s world .
one could say that the fact of its ubiquity is as clear as window glass .
because silica is an excellent insulator and can be easily created through thermal oxidation of the silicon substrate , sio@xmath1 is also the insulator of choice in the semiconductor industry .
optical fibers made from pure silica are widely used by the telecommunications industry and , because silica and silicates make up over 90% of the earth s crust , sio@xmath1 plays a major role in the geosciences .
liquid silica is the extreme case of a `` strong '' liquid . when cooled , its viscosity approaches the glass transition slowly , following the arrhenius law @xmath2 .
in contrast , the so - called `` fragile '' liquids reach this glass transition far more quickly . glasses rich in silica , but modified by other oxides to lower their viscosities , are `` strong '' liquids that have slow vitrification so are preferred by glassblowers who need time to work their magic .
simulations have indicated that liquid silica does not behave like a strong liquid for all temperatures , however . using the bks model @xcite ( see appendix a ) , vollmayr _ et al . _ found that at very high temperatures the diffusion greatly deviates from the arrhenius law ( and thus behaves like a fragile liquid ) , and that the temperature - dependence of the diffusion better fits the vogel - fulcher law @xcite .
it was later shown by horbach and kob @xcite that the temperature - dependence can also be fitted well by a power law of the shape @xmath3 in which the exponent @xmath4 is close to 2.1 ( compared to 1.4 for water ) and @xmath5 k. this temperature dependence is often found in simple liquids and has been described in terms of mode - coupling theory ( mct ) @xcite . a deviation from the arrhenius law has also been measured in other models of silica @xcite , and small deviations from a pure arrhenius law were found for the viscosity in experimental data @xcite .
this transition from fragile to strong upon cooling ( often called the `` fragile - to - strong crossover '' ) has also been found in simulations of other tetrahedral liquids , such as bef@xmath1 @xcite , silicon @xcite , and water @xcite .
this phenomenon is not restricted to tetrahedral liquids , however .
for example , it has been proposed that the fragile - to - strong crossover might be a behavior common to all metallic glass - forming liquids @xcite .
in addition to the fragile - to - strong crossover , it has been proposed that liquid silica also has a liquid - liquid critical point ( llcp ) @xcite much like that proposed for liquid water @xcite .
these phenomena may be related .
it was recently shown that in analog plastic crystal systems many strong glass - formers are accompanied by a singularity ( a lambda - type order - disorder transition ) at high temperatures , and that in silica this singularity could be a llcp @xcite .
the fragile - to - strong crossover arises simultaneously with a large increase of the isobaric heat capacity @xmath6 .
if a llcp exists in silica , this heat capacity maximum should have its origin in its critical fluctuations .
the discovery of a llcp in liquid silica would thus provide a unifying thermodynamic explanation for the behavior of liquid silica .
we consider here two different models of silica , the bks model by van beest _
_ @xcite and the wac model ( also known as the trim model for silica ) introduced by woodcock _
both models represent sio@xmath1 as a simple 1:2 mixture of si ions and o ions , i.e. , without any explicit bonds .
one difference between the two models is that wac uses full formal charges while in bks partial charges are used . for a detailed description of both models ,
see appendix a. all simulations are done using gromacs 4.6.1 @xcite , with @xmath7 ions , using the ewald sum ( pme ) for electrostatics , and the v - rescale thermostat @xcite to keep the temperature constant .
most simulations are done in the constant - volume / constant - temperature ( @xmath8 ) ensemble .
for the few constant - pressure ( @xmath9 ) simulations we use the parrinello - rahman barostat . for most of the simulations
we use a time step of 1 fs , but at very low temperatures we increase the time step to 4 fs to speed up the simulations to approximately 250 ns / day .
we carefully check the temperatures below which the 4 fs time step gives the same results as the 1 fs time step and do not include any 4 fs data that lead to a small difference in pressure , energy , or diffusion . as a measure of the equilibration time
, we define @xmath10 as the time at which @xmath11 nm , i.e. , the average time it requires for an o ion to move twice its diameter of 0.28 nm . most simulations run for over @xmath12 , well beyond the time necessary for the system to reach equilibrium . for the range of temperatures and pressures considered here , the root mean squared displacement of the o ion is roughly 1.11.6 times that of the si ion , this factor being the largest at low temperatures and low pressures .
an important structural feature is the coordination number of si by o , since a tetrahedral network is defined by 4-coordination of the network centers .
we calculate the si coordination number by the usual method , integrating the si - o radial distribution function up to the first minimum . for both models , and at all state points considered here ( below 10 gpa ) , the coordination number lies between 4.0 and 4.9 .
the coordination number is the largest at high densities , and levels off to 4 when the density is decreased and the pressure reaches zero and becomes negative .
the most direct method of locating a critical point is to calculate the pressure @xmath13 as a function of temperature @xmath14 along different isochores . in a @xmath15-diagram
the isochores cross within the coexistence region and at the critical point . at those state points
( at a given @xmath13 and @xmath14 ) the system is a combination of two different phases with different densities .
one can also locate a critical point by plotting the isotherms in a @xmath16-diagram in order to determine the region in which the slope of the isotherms becomes zero ( critical point ) or negative ( coexistence region ) . because it is easier to determine whether two lines are crossing than whether a curve is flat , we study the isochores .
figure [ fig : isochores_bks_wac ] shows the @xmath15-diagrams with the isochores of bks and wac .
obtain a negative curvature .
if this curvature becomes more negative as @xmath14 goes down , then it is possible that the isochores will not cross below 3500 k. we conclude that for the temperatures currently accessible , the isochores alone are insufficient to demonstrate a llcp in wac .
, title="fig : " ] + obtain a negative curvature .
if this curvature becomes more negative as @xmath14 goes down , then it is possible that the isochores will not cross below 3500 k. we conclude that for the temperatures currently accessible , the isochores alone are insufficient to demonstrate a llcp in wac .
, title="fig : " ] both diagrams are similar . there is a clear density anomaly to the left of the temperature of maximum density ( tmd ) , and if we raise the temperature by approximately 4000 k then the bks isochores match those of wac reasonably well .
thus , based on the isochores in fig .
[ fig : isochores_bks_wac ] , one could say that bks and wac are very similar systems , and that they mainly differ in a shift of temperature . at very low @xmath13 and high @xmath14 the liquid phase is bound by the liquid - gas ( or liquid - vacuum )
spinodal , and lowering @xmath13 below the spinodal leads to spontaneous bubble formation . at very low @xmath14
the liquid becomes a glass , and the diffusion coefficient drops rapidly . because the time it takes to equilibrate the system is inversely proportional to the rate of diffusion ,
simulations require too much time once the oxygen diffusion @xmath17 drops below @xmath18 @xmath19/s , which is where the isochores stop in fig .
[ fig : isochores_bks_wac ] . for both models
this limit is reached at a higher temperature for low @xmath13 than for high @xmath13 .
this is caused by the diffusion anomaly ( an increase in @xmath13 leads to an _ increase _ in diffusion ) , which is present in both bks and wac models .
no crystallization was observed , unless the pressure was raised to values far outside the range of our detailed studies ( e.g. , above 40 gpa the wac liquid spontaneously crystallizes into an 8-coordinated crystal ) .
normally , crystallization is readily detected by a rapid drift of the energy to lower values .
however , when the diffusivity is very low ( as in the present system , in the domain of greatest interest ) the situation is different and crystal growth can be unobservably slow .
more direct tests are then needed . in the present case
we have sought information on crystal growth and melting by creating a crystal front ( half simulation box of the liquid interfacing with half box of the topologically closest crystal ) and have watched the crystal front receding at high temperature .
however , the attempt to determine melting point by lowering the temperature and observing reversal of the interface motion , was unsuccessful because the growth rate became unobservably small ( observed over microseconds ) before any reversal was seen .
we conclude that , since this crystal front was put in by hand , the possibility of crystallization by _ spontaneous _ nucleation ( always the slowest step ) followed by crystal growth , is zero . based on the fitting and extrapolation of data ,
previous studies have predicted a liquid - liquid critical point ( llcp ) in both wac and bks @xcite . with the increase in computing power , and using the techniques to speed up the simulations discussed in sec . [
sec : methods ] , we are able to obtain data at lower temperatures than was previously possible . our results for bks ( fig .
[ fig : isochores_bks_wac]a ) show that for @xmath20 k the isochores are nearly parallel , and therefore a llcp in bks is very unlikely . on the other hand , the isochores of the wac model ( fig .
[ fig : isochores_bks_wac]b ) show a more interesting behavior in that they clearly approach one another at low @xmath14 in the vicinity of @xmath21 gpa .
if we only consider the wac isochores above 4000 k , then extrapolation would predict that the isochores cross around 3500 k and 5 gpa .
however , below 4000 k we see that the isochores are starting to display a negative curvature in the @xmath15-plane .
this signals an approach to a density minimum , which is the low-@xmath14 boundary of the density anomaly region .
the negative curvature makes it hard to perform an extrapolation that convincingly shows that the isochores cross at lower @xmath14 .
we can therefore only conclude that ( for the temperatures currently accessible ) the isochores are insufficient to prove or disprove the existence of a llcp in wac .
upon approaching a critical point , the response functions should diverge .
although true divergence occurs only in the thermodynamic limit @xmath22 , a large maximum should still be visible in response functions such as the isothermal compressibility @xmath23 and the isobaric heat capacity @xmath6 even when the box size is relatively small .
calculations using the ising model and finite size scaling techniques applied to simulation results have shown that ( for sufficiently large boxes ) the location of the critical point is very close to where both @xmath23 and @xmath6 reach their global maximum @xcite .
if a llcp truly exists in wac , then the @xmath15-diagrams of @xmath6 and @xmath23 should show a large @xmath6 maximum close to where @xmath23 has a maximum exactly where the isochores come together and where the llcp has been predicted to be .
( a ) ( b ) + ( c ) ( d ) figure [ fig : response_functions_wac ] shows four response functions for wac : ( a ) the isothermal compressibility @xmath23 , ( b ) the isobaric heat capacity @xmath6 , ( c ) the isobaric thermal expansivity @xmath24 , and ( d ) the isochoric heat capacity @xmath25 .
these have been obtained using @xmath8 simulations together with the smooth surface technique described in appendix [ appendix_smooth_surface ] . to check the results generated by this technique , we determine whether the response functions satisfy the thermodynamic relation @xmath26 . because of statistical errors in the data we find slight deviations from zero , but these are less than 1 j/(mol k ) in magnitude .
the compressibility @xmath23 in fig .
[ fig : response_functions_wac]a shows a clear global maximum near @xmath21 gpa and @xmath27 k , because this is where the isochores in fig .
[ fig : isochores_bks_wac]b are the closest together in terms of pressure .
it is quite likely that below 4000 k this maximum increases further .
if wac has a llcp then @xmath6 should also have a maximum in that vicinity .
however , fig .
[ fig : response_functions_wac]b shows that this is not the case .
there is clear global @xmath6 maximum , but it is located near @xmath28 gpa and @xmath29 k , which is far from the global @xmath23 maximum . therefore , based on the response functions , we conclude that wac does not have a llcp .
the isobaric thermal expansivity @xmath24 ( fig .
[ fig : response_functions_wac]c ) has a global minimum between the global maxima of @xmath6 and @xmath23 ( figs .
[ fig : response_functions_wac]a , b ) .
this should come as no surprise , since @xmath30 arises from fluctuations in entropy and @xmath31 from volume fluctuations , while the expansivity @xmath32 arises from a combination of both .
even though the global maxima occur at different places , the slopes @xmath33 of the loci of local maxima are the same , so it seems likely they have a common origin . because the system is not quite critical , the enthalpy fluctuations that determine the heat capacity can be statistically independent of the density fluctuations .
the variation of the heat capacity with temperature at constant pressure is shown over the temperature range in which the system remains in equilibrium , in fig .
[ fig : smoothingspline_bks_wac ] . fig .
[ fig : smoothingspline_bks_wac]b is basically a cross - section of fig .
[ fig : response_functions_wac]b .
we note first that at moderately high pressures , 8 gpa , there is no difference between the wac and bks models . in each case
the heat capacity reaches about 35 j/(k mol ) before the diffusion becomes too slow that we can no longer equilibrate .
this is 1.4 times the vibrational heat capacity of @xmath34 j/(k mol ) , as is typical of moderately fragile inorganic liquids ( e.g. anorthite , zncl@xmath1 ) right before ergodicity is broken @xcite .
however , at pressures between zero and 5 gpa , a major difference is seen between the models . near the tmd we have @xmath35 ( because the expansivity is very small ) so we can compare data with @xmath25 from scheidler _ et al . _
@xcite for the case of bks at @xmath36 .
the agreement is quantitative , up to the point where the earlier study was cut off .
our data confirms the existence of a peak in the equilibrium heat capacity an unusual behavior that was not reported in ref .
@xcite but had been noted in the earlier study of saika - voivod _ et al .
_ @xcite and was emphasized in ref .
@xcite .
although bks is far from having a critical point , the existence of this @xmath25 maximum reveals the tendency of this system which accords well with many aspects of experimental silica to develop the same anomalous entropy fluctuations , and an analog of the widom line made famous by water models . for the wac model ( which approaches criticality much more closely than bks does , as we have already seen in fig .
[ fig : isochores_bks_wac ] ) , this heat capacity peak becomes much more prominent , reminiscent of the behavior of the jagla model near its critical point .
@xmath6 reaches a value almost twice that of the vibrational component ; behavior unseen in any previous inorganic system except for bef@xmath1 which is a wac silica analog @xcite . at constant @xmath13 , followed by taking its derivative ( a slightly different method than was used in fig . [ fig : response_functions_wac]b ) . at 8
gpa there is no significant difference between the wac and bks models , but below 5 gpa wac has a large maximum in the range 50008000 k ( also clearly visible in fig .
[ fig : response_functions_wac]b ) . in panel
b we have included @xmath25 data of scheidler _ et al . _
@xcite ( red diamonds ) , which shows a maximum around 4500 k. near the tmd ( around 5000 k for @xmath36 ) the expansivity is small , which means that @xmath37 , in agreement with our results . for bks this
maximum is less clear in @xmath6 , though still visible .
because of small fluctuations in the data , it is difficult to obtain a fit of @xmath38 that produces a perfect estimate of @xmath39 , leading to artificial oscillations in @xmath6 . a larger data set would reduce this artifact . in addition , the smoothing spline method assumes zero curvature at the end - points of the data , and this leads to artifacts at very low @xmath14 and very high @xmath14 . for clarity
, we have removed the parts of the curves below the temperature at which @xmath6 starts to bend toward a constant a @xmath6 value .
, title="fig : " ] at constant @xmath13 , followed by taking its derivative ( a slightly different method than was used in fig . [ fig : response_functions_wac]b ) . at 8 gpa
there is no significant difference between the wac and bks models , but below 5 gpa wac has a large maximum in the range 50008000 k ( also clearly visible in fig .
[ fig : response_functions_wac]b ) . in panel
b we have included @xmath25 data of scheidler _ et al . _
@xcite ( red diamonds ) , which shows a maximum around 4500 k. near the tmd ( around 5000 k for @xmath36 ) the expansivity is small , which means that @xmath37 , in agreement with our results . for bks this
maximum is less clear in @xmath6 , though still visible .
because of small fluctuations in the data , it is difficult to obtain a fit of @xmath38 that produces a perfect estimate of @xmath39 , leading to artificial oscillations in @xmath6 . a larger data set would reduce this artifact . in addition , the smoothing spline method assumes zero curvature at the end - points of the data , and this leads to artifacts at very low @xmath14 and very high @xmath14 . for clarity , we have removed the parts of the curves below the temperature at which @xmath6 starts to bend toward a constant a @xmath6 value .
, title="fig : " ]
we find no llcp in either model within the accessible temperature range , although it is closely approached in the case of the wac potential near 4000 k and 5 gpa .
the isochores of bks , which are the most direct indicators of criticality in a physical system , fail to converge into a critical point . in the case of
wac we can not conclude anything from the isochores , but an analysis of the global extrema of the response functions indicates that there is no llcp in wac because the global @xmath6 maximum and the global @xmath23 maximum are significantly separated in the @xmath15-plane .
liquid silica forms a tetrahedral network of bonds , and below we will show that the lack of a llcp is related to the openness of this network structure , which in turn is related to the stiffness of the inter - tetrahedral bond angles .
in addition we will argue that criticality in wac could be achieved with an adaptation of the pair potential .
the occurrence of a llcp requires two competing liquid structures that can be in a ( meta - stable ) equilibrium with each other . in the case of a tetrahedral network - forming liquid
the two relevant structures are usually : ( i ) a high - density collapsed structure that is highly diffusive , and ( ii ) a low - density open network structure that is more rigid , i.e. , one that is still a liquid but less diffusive and more structured . because the high - density structure occupies a smaller volume but has higher entropy ( more disorder ) , the competition between these two structures is accompanied by a region with a density anomaly : @xmath40 .
the high - density structure is very stable and is the dominant structure at high temperatures , but the low - density structure requires a more delicate balance of forces in order to be stable .
if the bonds in the liquid are too flexible , the liquid collapses into the high - density structure . on the other hand ,
if the bonds are too rigid the liquid can no longer flow and becomes a glass .
there are several studies that address this situation .
the 2006 study of molinero _ et al . _
@xcite shows how reducing the three - body repulsion parameter @xmath41 in the stillinger - weber potential @xcite ( which controls the bond angle stiffness ) causes the first order liquid - liquid phase transition of silicon ( @xmath42 ) to disappear at @xmath36 when @xmath43 ( see fig .
[ fig : modsw ] ) .
this transition occurs between a low - density liquid and a high - density liquid , where both liquids are metastable with respect to the diamond cubic ( dc ) crystal .
crystallization to the dc crystal always occurs from the low - density liquid .
when @xmath44 crystallization happens so fast that it is no longer possible to accurately determine the temperature @xmath45 at which the phase transition occurs for @xmath36 .
simulations of the stillinger - weber model indicate that the llcp for @xmath46 is located at @xmath47 gpa and 1120 k @xcite .
since each value of @xmath41 defines a unique system with a unique critical pressure , the vanishing of the liquid - liquid transition at @xmath48 implies that this is the @xmath41 value for which the llcp is at @xmath36 .
isochore - crossing studies conducted elsewhere @xcite show that this is indeed the case , with @xmath49 k for @xmath50 .
it is clear that decreasing @xmath41 means decreasing the tetrahedrality and increasing density .
when @xmath48 the llcp shifts to positive pressures , and therefore the phase transition line can no longer be seen in fig .
[ fig : modsw ] , as it only considers @xmath36 .
we thus lack the information to determine exactly for which @xmath41 there is no llcp at _ any _ pressure , but it is certain that this happens at some value @xmath51 , since in the most extreme case of @xmath52 we are left with a simple lennard - jones - like model that has no llcp . and
temperature @xmath14 , at zero pressure @xcite .
the black triangles indicate the melting line of the diamond cubic ( dc ) crystal , while the green squares denote the melting line of the bcc crystal .
the dashed line separates the dc and bcc regions .
yellow circles indicate the transition temperature @xmath53 at which the liquid - liquid phase transition line crosses the @xmath36 isobar for that particular value of @xmath41 .
silicon is represented by @xmath46 and has a liquid - liquid critical point at @xmath47 gpa @xcite , and therefore all llcps for @xmath54 lie at negative pressures ( there is a llcp for each value of @xmath41 ) . for @xmath48
the llcps are at positive pressures and therefore the phase transition line can no longer be seen in this diagram .
when @xmath41 is large the system easily crystallizes , and therefore the phase transition line at @xmath36 can no longer be accurately located when @xmath44 . ] that weakening the tetrahedrality ( i.e. , making the tetrahedral bonds more flexible ) leads to the removal of a llcp , was also shown in 2012 by tu and co - authors using a different monatomic model @xcite .
the hamiltonian of this model includes a term that lowers the energy when particles are aligned along near - tetrahedral angles and thus favors a diamond cubic ground state .
the study of ref .
@xcite considers two versions : one that allows broad flexibility of the inter - tetrahedral bond angles ( leading to weak tetrahedrality ) , and another in which the bond angle is more constrained ( giving rise to strong tetrahedrality ) . the behavior for strong tetrahedrality is shown in fig .
[ fig : isochores_strong_tu ] , and we see that the isochores converge into a critical point .
if the tetrahedrality is weakened slightly , then the isochores separate , the llcp disappears , and the diagram starts to resemble that of fig . [ fig : isochores_bks_wac]b for wac .
it should be mentioned that a separation of the global @xmath6 and @xmath23 maxima also occurs in the weak tetrahedrality version ( as is the case for wac ) , while the @xmath6 and @xmath23 maxima are close together and near the llcp in the strong version of the model .
b ) show no llcp but closely resembles that of the strong tu model .
we can interpret this as that wac is _ close _ to having a llcp , but not close enough . if we were to enhance the tetrahedrality of wac , it is likely a llcp would appear . ]
finally we should consider the simulations done on `` patchy '' colloids by sciortino and coworkers . using the kern - frenkel ( kf ) model @xcite ( which consists of particles with tetrahedrally arranged sticky points ) , these authors demonstrated that the colloids developed tetrahedral network topologies , with each particle being surrounded by four others which is not itself surprising .
more interesting was the finding that , when the effective sizes of the patches were varied , conditions could be found in which not only were the relaxation kinetics strictly arrhenius in form , but also the amorphous state became the free energy ground state of the system , over a wide range of densities @xcite .
this corresponds to a more dramatic stabilization of the amorphous state than the kinetic stability observed in our work .
it signifies an absolute stability against crystallization on any time scale , i.e. , the system has become an `` ideal glassformer '' @xcite .
studies with the kf model have also demonstrated that highly directional bonds are needed to observe spontaneous crystallization in tetrahedral interacting particles @xcite , in agreement with the results found by molinero _
_ using the stillinger - weber family of potentials . since the kf colloids can be used to describe different tetrahedral models , they promote our understanding of tetrahedral liquids such as st2 and mw water , stilling - weber silicon , and bks silica .
surprisingly , there exists a mapping from these models to the kf model , using only a single parameter : the patch width @xcite .
the patch width is related to the flexibility of the bonds between the particles , and it is therefore likely that spontaneous crystallization and the existence of a llcp are related to bond angle flexibility .
all of these studies show that the occurrence of a llcp becomes less likely when the parameters controlling tetrahedrality are weakened .
unfortunately , the bks and wac models do not have an explicit parameter that controls tetrahedrality , such as the parameter @xmath41 in the stillinger - weber model . in this model
there is a direct relation between the value of @xmath41 and the tetrahedrality of the liquid measured by the orientational order parameter @xmath55 as defined by errington and debenedetti @xcite .
this parameter is constructed such that its average value @xmath56 will equal zero if all atoms are randomly distributed within the liquid , while @xmath57 for each atom within a perfect tetrahedral network ( such as in a cubic diamond lattice ) . for silica
the situation is more complicated .
it is not immediately clear how to define the tetrahedrality of a system that consists of two types of atoms .
one way would be to find for each si atom its four nearest neighboring si atoms and compute @xmath56 for this subset of atoms .
however , this measure would completely ignore the positions of the o atoms which form ionic bridges between the si atoms .
since the o - si - o bond angle deviates very little from the perfect tetrahedral angle of @xmath58 @xcite , it makes sense to focus on the inter - tetrahedral si - o - si bond angle instead .
it is commonly agreed that structures such as diamond cubic have maximum tetrahedrality , and for silica this corresponds to a system where all si - o - si bond angles are equal to @xmath59 ( such as @xmath0-cristobalite ) .
how much the inter - tetrahedral si - o - si bond angles differ from @xmath59 can thus be employed as a measure of the tetrahedrality , and we have therefore calculated this bond angle distribution for both bks and wac .
the location of the maximum in the si - o - si bond angle distribution ( i.e. , the most probable angle ) is a parameter that one could use to quantify the tetrahedrality .
if we denote the most probable angle at the lowest accessible temperature ( @xmath60 ) as @xmath61 , then the tetrahedrality parameter @xmath62 can be defined as @xmath63 , where @xmath64 . since the `` openness '' of the structure will increase with the average si - o - si angle
, one could also define the tetrahedrality using the volume ratio , i.e. , @xmath65 , which would require much less effort to calculate . here
@xmath66 is the volume of the perfect diamond cubic and @xmath67 is the system volume at some corresponding state , for instance at the tmd ( which is less arbitrary than @xmath60 ) .
let us consider the angular relations and the mechanical forces that determine them in more detail . in terms of the familiar ball - and - stick model , the si - o - si bond
could be represented by two sticks connected at the oxygen atom , with a spring in between the sticks .
this spring constrains the bond angle to some preferred bond angle @xmath68 , while the value of its spring constant @xmath69 ( the _ stiffness _ ) dictates how flexible the bond angle is . from the bond angle probability distribution @xmath70 , it is possible to estimate the values of the preferred bond angle @xmath68 and the bond angle stiffness @xmath69 . to extract the si - o - si bond angles from the data ,
we consider each o ion together with its two nearest si neighbors and calculate the angle between the two si - o bonds . in fig .
[ fig : bond_angles ] we show the resulting probability distributions @xmath70 of the si - o - si angle @xmath71 for bks and wac at zero pressure .
these curves have been measured before in previous studies @xcite but with less detail . as the temperature decreases , the width of the distribution decreases and the maximum shifts toward @xmath59 .
this implies that the liquid becomes more structured and stiffer .
this is to be expected , since at a high temperatures there are more thermal fluctuations and therefore @xmath70 has a broader distribution . in liquid silica for ( a ) the bks model and ( b ) the wac model .
as @xmath14 goes down , the most probable angle moves closer to @xmath59 while simultaneously the width of the distribution decreases .
the first phenomenon causes the liquid to expand upon cooling , while a reduction in width means that the bonds become stiffer , which leads to a decrease in diffusion .
both phenomena are related ( see below ) and are much stronger for wac than for bks . instead of @xmath70 it is better to consider @xmath72 , since a completely random distribution such as in the vapor has @xmath73 while @xmath74 is uniform ( see inset of panel a ) . for both models and all temperatures
@xmath74 resembles a normal distribution with mean @xmath59 .
this indicates that the preferred angle is in fact @xmath59 , and that the width of @xmath74 determines both the location of the peak in @xmath70 as well as its width .
, title="fig : " ] + in liquid silica for ( a ) the bks model and ( b ) the wac model . as @xmath14 goes down , the most probable angle moves closer to @xmath59 while simultaneously the width of the distribution decreases .
the first phenomenon causes the liquid to expand upon cooling , while a reduction in width means that the bonds become stiffer , which leads to a decrease in diffusion .
both phenomena are related ( see below ) and are much stronger for wac than for bks . instead of @xmath70 it is better to consider @xmath72 , since a completely random distribution such as in the vapor has @xmath73 while @xmath74 is uniform ( see inset of panel a ) . for both models and all temperatures
@xmath74 resembles a normal distribution with mean @xmath59 .
this indicates that the preferred angle is in fact @xmath59 , and that the width of @xmath74 determines both the location of the peak in @xmath70 as well as its width .
, title="fig : " ] plotting @xmath70 may not be the best way of presenting the bond angle distribution , as this distribution is biased toward @xmath75 angles .
this is particularly clear from the distribution of the vapor ( the thin black line in fig .
[ fig : bond_angles]a ) .
the ions in the vapor have no preferred position with respect to their neighbors , yet @xmath70 is not uniform but proportional to @xmath76 .
this is related to the fact that the infinitesimal area element of the unit sphere is @xmath77 rather than @xmath78 . as @xmath79 the area element @xmath80 approaches zero , and therefore @xmath81 at @xmath82 .
instead of @xmath70 it is better to consider the probability distribution @xmath83 , as is shown in the insets of fig .
[ fig : bond_angles ] .
the @xmath74 distribution of the vapor is a uniform distribution ( inset of fig .
[ fig : bond_angles]a ) . for the liquid ,
the distribution @xmath74 is approximately a normal distribution with its mean at @xmath84 .
evidently the most probable inter - tetrahedral angle ( the location of the @xmath70-peak ) is purely an effect of the width of this normal distribution combined with the fact that @xmath85 .
it is possible to interpret the bond angle distribution in terms of an effective potential @xmath86 , assuming that @xmath87 $ ] .
when the effective potential is harmonic , i.e. @xmath88 , the resulting probability distribution is a normal distribution with mean @xmath68 and a width that depends on temperature @xmath14 and stiffness @xmath69 . in general the effective potential will not be perfectly harmonic and includes anharmonic terms . because @xmath89 is an even function about @xmath90 , it is required that @xmath74 is as well , and therefore also @xmath86 . consequently , the leading - order anharmonic term in @xmath86 is of the fourth order .
the si - o - si bond angle distribution can thus be described by @xmath91 \label{eq : ptheta}\end{aligned}\ ] ] with @xmath92 a taylor series about the mean angle @xmath93 , @xmath94 here @xmath95 is a temperature - dependent normalization constant that ensures that the total probability @xmath96 is equal to one , and @xmath97 is the boltzmann constant .
the probability distributions of fig .
[ fig : bond_angles ] can be fitted quite well with eqs .
[ eq : ptheta ] and [ eq : ueff ] , even when the sixth power and higher - order terms are ignored .
the resulting values for the stiffness @xmath69 are shown in fig .
[ fig : k2 ] .
it is immediately clear that wac is far more rigid than bks . for bks the stiffness does not vary much with temperature ,
while increasing the pressure makes the bonds slightly less stiff .
the same is true for wac at high @xmath14 , but below 5 gpa the stiffness shows an increase when the liquid is cooled .
this increase is exactly where @xmath6 has its maximum in fig .
[ fig : response_functions_wac]b , and thus we may argue that the increase in @xmath6 is due to a structural change , namely the stiffening of the tetrahedral network . goes down with increasing pressure .
it is clear that bks has more flexible bonds ( small @xmath69 ) , and that wac is more rigid ( large @xmath69 ) and therefore `` more tetrahedral '' .
in addition wac shows a transition at low @xmath14 for @xmath98 gpa to a state with an even higher stiffness . ] from the isochores in fig .
[ fig : isochores_bks_wac]b it is clear that wac is very close to having a llcp .
if we compare the results of previous studies done on tetrahedral liquids @xcite with our results for bks and wac , then we see that the tetrahedrality of bks is far too small ( i.e. , the inter - tetrahedral bond angles are not sufficiently stiff ) to have a llcp , and that wac is close , but not close enough .
however , it might be possible to make a small change to the wac potential to enhance its tetrahedrality . one simple way to achieve this would be to add a repulsive term similar to the three - body interaction of the stillinger - weber model . this term should penalize any si - o - si configuration with an angle less than @xmath59 with a repulsive energy determined by the intensity parameter @xmath41 and the size of the deviation . the @xmath41 value associated with this interaction should be carefully chosen ; if @xmath41 is too small no llcp will arise , while applying a @xmath41 that is too large will likely lead to crystallization into a diamond ( @xmath0-cristobalite ) structure . it would be interesting to see at what value of @xmath69 this criticality is introduced , and if this value is the same across other tetrahedral models as well , but this is beyond the scope of the present project .
the results presented here are also relevant to the possible existence of a llcp in different water models , and highlight the importance of a thorough analysis of the o - h - o bond angle distribution .
such an analysis , possibly with the use of a bond angle stiffness parameter such as @xmath69 , might be able to predict if a particular water model will have a llcp .
unfortunately , to the best of our knowledge , it is currently not possible to measure these angles directly in experiments , as significant help from computer simulations is required to obtain the angular structure of liquid water @xcite .
although it has been suggested , based on a combination of simulation and theoretical considerations @xcite , that both bks and wac have llcps at temperatures beyond the accessible simulation range , our study suggests that neither bks nor wac can reach a critical point .
we have compared our results to those of other tetrahedral models @xcite , analyzed the bond angle distributions , and conclude that the lack of a llcp in both bks and wac is due to a lack of stiffness in the inter - tetrahedral si - o - si bond angles .
wac is close to criticality , but bks shows little sign of a llcp , and since the latter is considered to be the more realistic model for experimental silica , we expect that no llcp occurs in real silica either .
however , this does not mean that manifestations of criticality can never be observed .
as chatterjee and debenedetti @xcite have shown theoretically , even a weak tendency toward criticality ( as in bks ) can be amplified into a liquid - liquid phase separation in a binary system .
indeed this notion has been exploited elsewhere @xcite to interpret the ( much - studied @xcite but incompletely understood ) splitting out of an almost pure sio@xmath1 phase from such simple systems as the na@xmath1o - sio@xmath1 and li@xmath1o - sio@xmath1 binary glasses during supercooling .
we would like to thank p. debenedetti , v. molinero , h. arago , and c. calero for the many valuable discussions .
el and he s thank the national science foundation ( nsf ) chemistry division for support ( grant no .
che 12 - 13217 ) svb thanks the dr .
bernard w. gamson computational science center at yeshiva college for support .
caa acknowledges the support of this research through the the national science foundation ( nsf ) experimental chemistry program under collaborative grant no .
che 12 - 13265 .
one of the simplest models for silica is the wac model introduced by l. v. woodcock , c. a. angell , and p. cheeseman @xcite .
the model is sometimes also known as the transferable ion model ( trim ) because its potential is rather general and can also be used to model other ionic liquids @xcite . in the wac model ,
the material consists of a 1:2 mixture of si@xmath99 and o@xmath100 ions , without any explicit bonds .
apart from the electrostatic force , the ions also interact with each other via an exponential term : @xmath101 % \label{eq : def_wac_1}\end{aligned}\ ] ] here the subscripts @xmath102 indicate the species of the two ions involved , @xmath103 the charge of each ion ( @xmath104 , @xmath105 ) , @xmath106 the number of outer shell electrons , and @xmath107 the size of each ion ( @xmath108 nm , @xmath109 nm ) . for wac silica the parameters @xmath110 and @xmath111 are the same for all pairs : @xmath112 perg @xmath113 kj / mol and @xmath114 nm@xmath115 @xcite .
the potential can also be written as @xmath116 with @xmath117 kj / mol , @xmath118 kj / mol , and @xmath119 kj / mol .
the second model that we consider here is bks .
currently one of the most popular models , the bks model was introduced by b. w. h. van beest , g. j. kramer , and r. a. van santen @xcite and is similar to wac .
silica is again modeled as a simple 1:2 mixture of si- and o - ions , without explicit bonds . to produce results that better match experiments and _ ab initio _
simulations , and to be able to effectively represent screening effects , the charges in bks are not integer values of @xmath120 but instead are given by @xmath121 and @xmath122 .
in addition to this , the bks potential also differs from the wac model in that it includes an attractive @xmath123 term : @xmath124 in bks there is no interaction between two si - ions apart from the electrostatics , i.e. @xmath125 . the parameters for the si - o pair are @xmath126 ev , @xmath127 @xmath115 , and @xmath128 ev@xmath129 . for the o - o interaction ,
the numbers are @xmath130 ev , @xmath131 @xmath115 , and @xmath132 ev@xmath129 .
although the bks model has been quite successful in simulations of quartz and amorphous silica , at temperatures above @xmath133k two ions can come very close , causing problems . as @xmath134 the bks potential diverges to @xmath135 and the two ions fuse together
a non - physical phenomenon that is an artifact of the model .
one way to solve this issue is by including an additional repulsive term at very small @xmath136 , e.g. , by adding a @xmath137 term @xcite .
when such a large power is used , however , a small time step is required to prevent large forces , which leads to much slower simulations .
because of this , we instead adjust the bks potential at small @xmath136 by adding a second - degree polynomial for @xmath138 . here
@xmath139 is the point at which the original bks force has an inflection , i.e. , where @xmath140 .
we choose the coefficients of the polynomial such that the new potential @xmath141 has no inflection at @xmath142 .
adding the polynomial still leads to @xmath143 when @xmath144 , but increases the height of the energy barrier sufficiently to allow us to simulate the high temperatures we wish to explore . choosing a short - range correction to bks has been found to have little effect on the simulation results , and merely prevents the ions from fusing . to further speed up the simulations , we modify the bks potential as described by k. vollmayr , w. kob , and k. binder in ref .
@xcite , and truncate and shift the potential at @xmath145 nm .
although this truncation leads to a shift in pressure , it otherwise produces approximately the same results @xcite . in conclusion ,
the modified bks potential we use is given by @xmath146 with the parameter values for @xmath147 and @xmath148 listed in table [ tab : parameters_mod_bks ] .
for the si - si interaction the potential is @xmath149 and does not involve any cutoffs , apart from the real - space cutoff of the ewald sum . .
parameters of the modified bks potential of eq .
( [ eq : def_mod_bks ] ) . because si - si only has the ( repulsive ) coulomb interaction , all parameters are zero for si - si .
one mol here indicates one mol of ions , not one mol of sio@xmath1 molecules . [ cols="<,>,>,<",options="header " , ]
in order to construct isobaric response functions from a large set of constant - volume ( @xmath8 ) data , some type of fit or interpolation is needed .
for example , to calculate @xmath150 we consider the enthalpy @xmath151 as a function of both @xmath13 and @xmath14 and fit the data @xmath152 $ ] with a smooth 3-dimensional surface @xmath153 .
abrupt changes in @xmath153 lead to large spikes in its derivative @xmath154 , and thus the @xmath153 surface must be smooth if we are to obtain a meaningful @xmath6 . fitting a surface rather than
a curve has the additional advantage that more data is taken into account , resulting in better statistics .
an alternative approach is to calculate @xmath6 via fluctuations in @xmath151 , but it has been shown @xcite that first fitting @xmath38 and then taking a derivative leads to cleaner results .
it is of course easier to calculate @xmath6 by doing constant - pressure ( @xmath9 ) simulations instead , but then one would have the same problem with calculating @xmath25 .
we conclude that we can easily calculate all response functions if we apply a smooth surface fit @xmath155 to a set of 3-dimensional points @xmath156 . fitting a surface to a set of points
means striking a balance between the `` smoothness '' of the fit and the fitting error induced .
one measure of smoothness is the laplacian @xmath157 , since a small laplacian means little change in the slope of @xmath155 , and thus a smoother function .
hence , to obtain a smooth surface fit @xmath155 through the data points @xmath156 with @xmath158 , we minimize @xmath159 ^ 2 + \iint \left| \nabla^2 f(x , y ) \right|^2 dx\,dy .
\label{eq : smooth_surface}\end{aligned}\ ] ] the weights @xmath160 provide the balance between the smoothness and the fitting error .
if we set @xmath160 too low , we obtain a very smooth fitting function @xmath155 that poorly represents the data .
if we set @xmath160 too high , the function @xmath155 will go through all the data points but will show large variations . because large variations in the surface lead to even larger variations in the derivatives , the @xmath153 surface must be very smooth when we calculate the @xmath6 . fortunately , introducing small fitting errors does not cause problems , because the simulation data already suffers from small statistical errors .
if the underlying response function is in fact smooth , then it is possible to use the fitting errors to partially cancel the statistical errors .
minimization of the functional @xmath161 in eq .
[ eq : smooth_surface ] is not a new concept . for example , the csaps function in matlab applies a similar minimization scheme to calculate a cubic smoothing spline . as opposed to this matlab function
, we do not impose the constraint that @xmath155 is a tensor product spline , but instead represent @xmath155 by a set of @xmath162 points @xmath163 placed on a regular grid @xmath164 .
bilinear interpolation is used to estimate the value of @xmath155 between these grid points , and the derivatives and the laplacian are calculated using finite ( central ) differences . to compensate for the reduced number of data points near the edges of the domain
, we recommend that higher - order differences near the edges be used . , liquids , freezing and the glass transition , in _ proceedings of the les houches summer school of theoretical physics , session li _ , edited by j .-
hansen , d. levesque , and j. zinn - justin , pp .
287503 , north - holland , amsterdam , 1991 , 1989 . and
c. a. angell , comparison of pair potential models for the simulation of liquid sio2 : thermodynamic , angular - distribution , and diffusional properties , in _ physics meets mineralogy : condensed matter physics in the geosciences _ , edited by h. aoki , y. syono , and r. j. hemley , chapter 6.1 , pp .
325339 , cambridge university press , cambridge , england , 2000 . and m. hemmati , glass transitions and critical points in orientationally disordered crystals and structural glassformers : ( `` strong '' liquids are more interesting than we thought ) , in _
4th international symposium on slow dynamics in complex systems _ , edited by m. tokuyama and i. oppenheimer , volume 1518 , p. 9
, aip conf .
proc . , 2013 . , t. a. kesselring , g. franzese , s. v. buldyrev , h. j. herrmann , and h. e. stanley , response functions near the liquid - liquid critical point of st2 water , in _
4th international symposium on slow dynamics in complex systems _ , edited by m. tokuyama and i. oppenheimer , volume 1518 , pp . 520526 , aip conf .
, 2013 .
, p. h. poole , and m. hemmati , a new interpretation of liquid - liquid unmixing in classical alkali silicate glasses , in _ proc .
12th east european glass conf .
( varna , bulgaria ) _ , edited by b. samunova and y. demetriew , pp . 100109 , 1996 . and b. g. varshal , causes of phase separation in simple silicate systems , in _ phase - separation phenomena in glasses _ , edited by e. a. porai - koshits , volume 8 of _ the structure of glass _ , pp
711 , consultants bureau , new york , 1973 . | previous research has indicated the possible existence of a liquid - liquid critical point ( llcp ) in models of silica at high pressure . to clarify this interesting question
we run extended molecular dynamics simulations of two different silica models ( wac and bks ) and perform a detailed analysis of the liquid at temperatures much lower than those previously simulated .
we find no llcp in either model within the accessible temperature range , although it is closely approached in the case of the wac potential near 4000 k and 5 gpa .
comparing our results with those obtained for other tetrahedral liquids , and relating the average si - o - si bond angle and liquid density at the model glass temperature to those of the ice - like @xmath0-cristobalite structure , we conclude that the absence of a critical point can be attributed to insufficient `` stiffness '' in the bond angle .
we hypothesize that a modification of the potential to mildly favor larger average bond angles will generate a llcp in a temperature range that is accessible to simulation .
the tendency to crystallize in these models is extremely weak in the pressure range studied , although this tendency will undoubtedly increase with increasing stiffness . |
one of the most interesting questions in young stellar object ( yso ) research is how their jets , e.g. @xcite , @xcite or @xcite , are collimated and accelerated . to address this problem
one must obtain not only high spatial resolution but in addition kinematic information as close as possible to their source .
ground - based long - slit spectroscopic studies of optically visible jet sources have shown that the structure and kinematics of the outflow region on scales @xmath51@xmath0 is rather complex ( e.g. @xcite ; @xcite ) . in a significant fraction of classical t tauri stars ( cttss )
the forbidden emission lines ( fels ) show two ( or more ) blueshifted velocity components which have very different properties ( note that in many cttss the corresponding redshifted part of the flow is occulted by a circumstellar disk , at least close to the source ) . in the case of double - peaked fel profiles ,
the emission consists of a so - called high - velocity component ( hvc ) , having typical radial velocities of -60 to -200 kms@xmath6 and a low - velocity component ( lvc ) , having typical radial velocities of -5 to -20 kms@xmath6 with respect to the systemic velocity of the star ( see @xcite and @xcite for comprehensive discussions ) .
while the hvc can be spatially very extended ( at least in the outflow direction ) and is often identified with a jet , the lvc is much more compact ( @xmath51@xmath0 ) .
other differences between the two components have also been noted , for example the lvc appears to be of higher density but lower excitation than the hvc @xcite and the lvc shows a clear inverse correlation of velocity with increasing critical density .
the latter effect has been interpreted as evidence of acceleration in the lvc with increasing distance from the star .
a number of theories have been put forward to explain the origin of the various fel components .
for example kwan and tademaru ( 1988 , 1995 ) have suggested that the lvc and hvc are separate flows . according to their model
the lvc is a poorly collimated wind coming from the outer periphery of the yso disk while the hvc is a separate jet launched from closer to the star .
others have sought to explain the observations in terms of a one component , non - isotropic , wind model in which the appearence of separate fel components is due to projection effects @xcite . in any event , @xcite has shown that the luminosity of the two components is tightly correlated , which seems to imply that they are not independent flows , and that the apparent dominance of one component over the other may be a density effect . in order to examine in more detail the nature of the compact fel region , we have observed the ctts dg tau with stis on - board the hst .
multiple overlapping slit positions parallel to the outflow from this star were chosen so as to build up a 3-d spatial intensity - velocity `` cube '' .
our target was picked not only because it is one of the closest cttss , but more significantly because its fel region has a broad range of velocities , probably also a result of the relatively small angle between the jet axis and the line of sight ( @xmath7 , see @xcite ) .
this is important given the moderate wavelength resolution of stis .
historically , dg tau was amongst the first t tauri stars from which a jet - like outflow ( hh 158 ) was discovered @xcite and it has been imaged by the hst prior to the installation of the telescope s correcting optics @xcite . on large scales ( @xmath8 10@xmath0 ) the jet seems to terminate in a bow shock @xcite while high resolution spectro - imaging ground - based studies @xcite have shown that the flow close to the star contains at least two resolved knots ( at 2.7@xmath0 and 4@xmath0 , epoch 1994.8 ) .
the outermost of these has a morphology and velocity gradient which is also consistent with it being a bow shock . in 2
we describe our observational technique and give details of our stis data .
our primary results are described in 3 and discussed in 4 .
a more detailed analysis will be presented in @xcite .
stis spectra of dg tau were taken on january 14 1999 using the g750 m grating and a central wavelength setting of 6581 . the spectral range of 562 included several strong forbidden lines ( [ oi]@xmath96300,6363 , [ nii]@xmath96548,6583 , [ sii]@xmath96716,6731 ) as well as h@xmath1 .
the stis / ccd detector , a 1024@xmath101024 pixel array , had a spectral resolution of 0.56pixel@xmath6 and a nominal sampling of 0.05@xmath0pixel@xmath3 in the dispersion and spatial directions respectively .
the actual spatial resolution , however , was limited by the psf of the hst in the red to a fwhm of approximately 0.1@xmath0 .
the slit aperture was [email protected] arcsec@xmath11 .
seven different long - slit spectra were taken , keeping the slit parallel to the outflow axis ( p.a .
226@xmath12 ) , but with steps of 0.07@xmath0 in the transverse direction , i.e. with offsets southeast and northwest of the jet axis . in this way we built - up a 3-d flux density / radial velocity data cube of the optical outflow from dg tau , with a total spatial width of about 0.5@xmath0 in the direction perpendicular to the jet axis .
the spectra are labelled s1,s2, ... s7 going from the southeast to the north - west ; the central slit position ( s4 ) coincided with the star ( ra 4@xmath13 27@xmath14 [email protected] , dec + 26@xmath12 6@xmath16 16.8@xmath0 . )
exposure times were split in order to facilitate removal of cosmic rays .
total exposure times were approximately 2740s for s1
s5 , 1930s for s6 and 2050s for s7 .
note that the somewhat shorter exposure times for the last two spectra were due to minor problems with the hst .
the pipeline spectra were found to be contaminated by a large number of `` hot '' pixels and so the raw data were first fully re - calibrated using the _ calstis _ suite of programs and new reference files made available by the stsci .
subsequent reduction and data analysis was carried out using standard iraf routines . in order to minimise contrast problems close to the source , the stellar continuum , and that of a faint reflection nebula extending up to 2@xmath0 from the source , were carefully subtracted from all images .
this operation proved to be particularly critical for the central spectra s3 , s4 , and s5 , due to the appearance of artificial undulations in the stellar continuum , caused by sampling effects ( see the stis handbook v3.0 ) . from the acquired spectra we created synthetic 2-d images for the various lines in four broad distinct radial velocity intervals ( figs .
these images were constructed by adjoining the seven row - averaged columns of pixels obtained from s1s7 for the corresponding velocity bin .
averaging was done over five columns , corresponding to approximately 2.83 , in the dispersion direction . in order to study the lowest velocity emission in finer detail , we also ( fig .
3 ) carried out a similar procedure in [ oi]@xmath176300 and [ sii]@xmath176731 but this time using just one column in the dispersion direction .
all velocities quoted in this _
letter _ are systemic using v@xmath18 + 16.5kms@xmath6 , as derived from the li@xmath176707 photospheric absorption line .
our composite images of the outflow close to dg tau in the four broad velocity bins , and in each of four lines ( h@xmath1 , [ nii]@xmath176583 , [ sii]@xmath176731 , and [ oi]@xmath176363 ) , are shown in figs . 1 and 2 .
note that we did not include [ oi]@xmath176300 in these figures as the blueward portion of this line was missing from our spectra due to a restricted choice for the stis central wavelength .
the velocity range is approximately + 50kms@xmath3 to -450kms@xmath3 and the four velocity bins ( low , medium , high , and very high ) , with widths of about 125kms@xmath3 , represent increasingly blueshifted velocities . on all figures
the position of the star ( i.e. the peak in the continuum subtracted light ) is marked .
similar images , for the [ oi]@xmath176300 and [ sii]@xmath176731 lines , but only for the lowest velocities ( from about + 60kms@xmath3 to -70kms@xmath3 ) and with bin widths of only about 25kms@xmath3 , are shown in fig .
first of all , we note that the appearence of the outflow changes , remarkably in some cases , from line to line and from velocity bin to velocity bin ( see in particular figs . 1 and 2 ) .
the jet from dg tau is , as one might expect , most evident at the highest velocities and is visible as a narrow feature up to about 0.7@xmath0 from the star
. we will discuss it in more depth shortly . aside from the jet a limb - brightened bubble - like structure can also be observed that extends between 0.4@xmath0 and 1.5@xmath0 from dg tau .
the `` bubble '' , which is most evident at intermediate velocities , in combination with the jet , gives an overall pear - shaped morphology to the emission .
comparison of figs . 1 , 2 and 3 lead to a number of important results : + ( i ) the lowest velocity emission close to the star ( @xmath50.3@xmath0 ) is spatially broad , i.e. it is well resolved in the transverse direction to the outflow , but it does not extend far along the flow .
+ ( ii ) within @xmath50.5@xmath0 from the star , there appears to be a more or less gradual increase in the degree of collimation with velocity : at high and very high velocities , the flow is primarily confined to the central axis .
+ ( iii ) beyond 0.5@xmath0 from the star there are at least two well - defined structures ( which we have labelled a1 and a2 in the medium velocity h@xmath1 image following the knot nomenclature of @xcite ) .
the outermost one , a1 , is bow - shaped and observed in all lines but the innermost one , a2 , is more easily seen in h@xmath1 .
there is very high velocity gas immediately behind a1 , which is seen in all lines but which is not symmetrically distributed around the jet axis .
the opposite effect is , incidentally , seen near a2 where the higher velocity gas appears in the _ downstream _ direction . the emission from a2
is also distributed asymmetrically with respect to the outflow axis .
+ ( iv ) comparison of the [ oi ] , [ sii ] , and [ nii ] data ( in figs .
1 and 2 ) show that the central jet can be traced much closer to the source in [ oi ] and [ nii ] than in [ sii ] .
this result is probably due to a combination of quenching and excitation effects ( see @xcite for details ) .
note that the [ oi ] line has a 100 ( 10 ) times higher critical density than the [ sii ] ( [ nii ] ) line , and that the [ nii ] line usually traces gas of higher excitation .
the emission peak ( centroid ) in the [ sii ] and [ oi ] images moves outwards with increasing velocity .
this effect is visible at all velocities , but it is most obvious at the highest velocities where virtually no [ sii ] emission and only faint [ oi ] emission is seen in the region @xmath51@xmath0 . since the high velocity jet can be traced right back to the source in h@xmath1 , the lack of high - velocity and the presence of low - velocity fel emission at @xmath19 1@xmath0 can probably only be explained by quenching effects in _ both _ the longitudinal and lateral directions . in other words
our data suggest that the jet density increases not only with proximity to the star ( in the longitudinal direction ) but also transversely toward the outflow axis . without the latter effect one would not be able to explain why
there is so much spatially extended and easily quenchable _ low - velocity _ fel emission seen so close to the source .
+ ( v ) quenching effects likewise manifest themselves at low velocities ( see fig .
3 ) in that the centroid of the low velocity [ oi]@xmath176300 emission is closer to the star than the centroid of the [ sii]@xmath176731 emission .
this was expected on the basis of groundbased measurements @xcite .
+ ( vi ) in the region close to the star ( @xmath50.7@xmath0 ) [ nii ] emission comes primarily from the high velocity jet implying this is the region with the highest excitation .
+ ( vii ) the medium velocity bow - shaped structure a1 is less evident in [ sii ] and [ oi ] than in h@xmath1 , suggesting that its excitation level may be high .
+ ( viii ) there is a strong low velocity peak in h@xmath1 emission coinciding with the star .
this is almost certainly scattered emission from a region close to dg tau , including possibly a magnetospheric contribution @xcite and therefore does not constitute part of the outflow _ per se_.
the stis observations presented here clearly show that the hvc emission in dg tau comes from the most highly focused part of the outflow i.e. the jet .
the jet at high velocities can be traced back to at least 0.1@xmath0 ( 15au ) from the star where quenching effects become important even in the case of the [ oi]@xmath96300,6363 lines , which have the highest critical density ( @xmath20 @xmath21 ) among the studied fels .
obviously it is likely that the jet is collimated on even smaller scales .
the shifts of the emission centroids in the different lines and in the different velocity bins can be explained quite naturally if there is an increase in jet density not only with proximity to the star ( in the longitudinal direction ) but also to the central outflow axis .
that is to say the high velocity `` core '' of the jet is denser than its periphery .
this is confirmed by an inspection of the [ sii]@xmath96716,6731 doublet ratio ( see fig . 2 and bacciotti et al .
note also that structures such as a1 and a2 are reminiscent of the `` bubbles '' recently seen by @xcite in the case of xz tauri .
these are almost certainly internal working surfaces caused by temporal variations in the outflow from dg tau .
certainly there is plenty of evidence for strong jet velocity variations in dg tau ( and in many other ctts stars ) on timescales of years .
for example , the data of @xcite and @xcite show an increase in the radial velocity of the jet at d @xmath4 0.5@xmath0 by a factor of 2 within 8 years .
also the proper motion data of @xcite indicate large velocity variations .
finally , several of the knots in the dg tau jet show bow shock - like structures ( see also the hst archive data presented in @xcite ) and these provide indirect evidence for strong velocity variations . + turning now to the lvc , its nature still remains somewhat enigmatic .
it was already clear from groundbased observations that the lvc and hvc differ in many properties such as density and excitation .
a new difference is reported here i.e. the rather large spatial width w of the lvc perpendicular to the jet at distances d of about [email protected]@xmath0 from the source .
a comparison of w ( at fwhm ) between the high and low - velocity emission in [ sii ] and [ oi ] shows that the average w(lvc ) @xmath4 0.18@xmath0 while the hvc is hardly resolved at a distance d of 0.2@xmath0 .
we note that this comparison of spatial widths can only be done using our [ sii ] and [ oi ] `` images '' as the h@xmath1 data is heavily contaminated by stellar h@xmath1 emission and in [ nii ] the lvc is very weak .
another interesting result of our study is the smaller velocity of the lvc at the edges of the flow ( for full details see @xcite ) .
for example at d=0.2@xmath0 the lvc peaks at @xmath4 -95kms@xmath3 in [ oi ] for the central slit position ( s4 ) while in the two outermost slit positions ( s1,s7 ) it peaks at about -18kms@xmath3 .
the corresponding values for the [ sii ] line are -60kms@xmath3 and -40kms@xmath3 , respectively .
since these lines are optically thin , such observations clearly point to a rise in the average lvc velocity as the central outflow axis is approached .
+ to what degree the observations of dg tau presented here , particularly of the lvc , are representative of other cttss is an open question .
dg tau is one of the most active cttss known and we caution that the lvc of dg tau is unusual in that it has the highest absolute velocity of all the cttss listed by @xcite , as well as one of the highest accretion rates .
that said it shares the typical properties of other lvcs and , in particular , the ratio of its luminosity ( l ) to that of the dg tau hvc is in perfect agreement with the l(hvc ) v. l(lvc ) relationship noted by @xcite .
thus dg tau may simply be displaying the higher activity tail of the distribution of outflow properties amongst cttss . in conclusion
our stis observations show , for the first time , a quasi - continuous variation in the outflow velocity close to a yso in the transverse direction to the flow .
detailed studies are required , however , to test whether these observations can constrain models for the generation of the lvc and hvc .
bacciotti , f. , ray , t.p . ,
mundt , r. , eislffel , j. , solf , j. and camenzind , m. 2000 , in preparation calvet , n. 1997 , in herbig - haro flows and the birth of low mass stars , iau symp .
182 , eds .
b. reipurth and c. bertout ( kluwer academic publishers ) , 417 camenzind , m. 1997 , in herbig - haro flows and the birth of low mass stars , iau symp .
182 , eds .
b. reipurth and c. bertout ( kluwer academic publishers ) , 241 edwards , s. 1997 , in herbig - haro flows and the birth of low mass stars , iau symp .
182 , eds .
b. reipurth and c. bertout ( kluwer academic publishers ) , 433 eislffel , j. , and mundt , r. 1997 , , 114 , 280 eislffel , j. , and mundt , r. 1998 , , 115 , 1554 hartigan , p. , edwards , s. , and ghandour , l. 1995 , , 452 , 736 hartmann , l. , and raymond , j.c .
1989 , , 337 , 903 hirth , g. , mundt , r. , and solf , j. 1994 , , 285 , 929 hirth , g. , mundt , r. , and solf , j. 1997 , , 126 , 437 kepner , j. , hartigan , p. , yang , c. , and strom , s. 1993 , , 415 , 119 kwan , j. , and tademaru , e. , 1995 , , 454 , 382 kwan , j. , and tademaru , e. , 1998 , , 332 , 41 krist , j.e . , et al . 1999 , , 515 , 35 lavalley , c. , cabrit , s. , dougados , c. , ferruit , p. , and bacon , r. 1997 , , 327 , 671 mundt , r. , brugel , e , and bhrke , t. 1997 , , 319 , 275 mundt , r. , and fried , j. , 1983 , , 274 , 83 ouyed , r. , and pudritz , r.e .
1994 , , 423 , 753 ray , t.p .
1996 , in the nato asi on solar and astrophysical mhd flows , ed .
k. tsinganos ( kluwer academic publishers ) , 539 safier , p.n .
1993 , , 408 , 148 solf , j. 1997 , in herbig - haro flows and the birth of low mass stars , iau sympos . 182 , eds .
b. reipurth and c. bertout ( kluwer academic publishers ) , 63 solf , j. , and bhm , k.h .
1993 , , 410 , 31 | we have carried out a spatio - kinematic study of the outflow from the classical t tauri star dg tau using the _ space telescope imaging spectrograph _ ( stis ) on board the hubble space telescope ( hst ) .
a series of seven spatially offset long - slit spectra spaced by 0.07@xmath0 were obtained along the axis of the outflow to build up a 3-d intensity - velocity `` cube '' in various forbidden emission lines ( fels ) and h@xmath1 .
here we present high spatial resolution synthetic line images close to the star in distinct radial velocity intervals ( from @xmath2 + 50kms@xmath3 to @xmath4 -450kms@xmath3 in four bins , each @xmath4 125kms@xmath3 wide ) . the lowest velocity emission is also examined in finer detail ( from + 60kms@xmath3 to -70kms@xmath3 in five bins @xmath4 25kms@xmath3 wide ) .
we have found that the highest velocity and most highly collimated component , i.e. the jet , can be traced from dg tau to a distance [email protected]@xmath0 .
the jet is on the axis of a pear - shaped limb - brightened bubble which extends between 0.4@xmath0 and 1.5@xmath0 from the source and which we interpret as a bow shock .
other condensations are seen close to the star indicating ongoing temporal variations in the flow .
the low - velocity component of the outflow is found to be spatially wide close to the source ( @xmath40.2@xmath0 at d=0.2@xmath0 ) , in contrast to the high velocity jet ( width @xmath50.1@xmath0 ) .
we have also found evidence to suggest that not only does the density in the outflow increase longitudinally with proximity to the source but that it also increases laterally towards the flow axis .
thus , at least in the case of dg tau , the flow becomes gradually denser as it increases in velocity and becomes more collimated .
our observations show a continous bracketing of the higher speed central flow within the lower speed , less collimated , broader flow , down to the lowest velocity scales .
this suggests that the low and high velocity fels in the highly active t tauri star dg tau are intimately related .
implications of these observations for fel models will be considered in a future paper @xcite . |
pre - main sequence ( pms ) stars are objects still contracting to the main sequence and generally surrounded by disks and/or envelopes of circumstellar dust and gas , which represent remnant material from their still - ongoing formation .
this circumstellar matter is responsible for phenomenons such as emission lines , p cygni profiles , ir and uv excesses , and polarization .
this polarization is produced by scattering on dust grains and has been known for a number of years ( see bastien 1996 for a review ) .
it is also well known that when a binary star ( of any evolutionary status ) is surrounded by circumstellar matter , its polarization varies as a function of the orbital period .
some models seek to reproduce these variations ( see for example rudy & kemp 1978 ; brown , mclean , & emslie 1978 ) in order to find the orbital inclination of these binary systems .
in particular , the work from brown et al .
( 1978 ) ( hereafter referred to as bme ) uses first and second - order fourier analysis of the stokes curves to give , in addition to the orbital inclination , moments related to the distribution of the scatterers in the circumstellar and circumbinary environment
. it could be very interesting to use the bme formalism to find the orbital inclination of pms binaries , since spectroscopic observations coupled with an orbital inclination can yield the absolute mass of each star .
these masses could then be compared with theoretical models of the formation of binary stars , and to masses derived from other types of observations ( photometry and theoretical evolutionary tracks for example ) .
however , the bme formalism was developed for thomson scattering in optically thin envelopes , and for binaries in circular orbits . since polarization in pms stars
is produced by scattering on dust grains , and most of the known pms spectroscopic binaries have eccentric orbits , the bme formalism can not be used a priori .
studies were undertaken to verify the applicability of the bme formalism for mie scattering and eccentric orbits ( see manset & bastien 2000 , 2001 , hereafter referred to as paper i and ii respectively ) and have shown that the bme analysis can still be applied in those cases , with a few limitations . in this context , we have obtained polarimetric observations of 24 pms spectroscopic binaries , following ( with @xmath1 observations ) the shortest - period ( @xmath2d ) and brightest ( @xmath3 ) ones .
the observations were used to study the general polarimetric variability of these systems , correlations between these variations and physical characteristics , periodicity in the variations , and to try to determine the orbital inclination . in this paper
, we present the observational method , method of determination of the interstellar polarization , analysis of the polarimetric variability and of the periodic variations , along with data for the only herbig ae / be star of our sample , mwc 1080 .
the observations and discussion for the rest of the sample will be presented in future papers .
polarimetric data were taken at the observatoire du mont mgantic ( omm ) , canada , with beauty and the beast , a two - channel photo - electric polarimeter used at the f/8 cassegrain focus of the 1.6 m telescope . in this polarimeter ,
the analyzer is a wollaston prism which is used in combination with a pockels cell acting as a variable quarter - wave plate , and an additional quarter - wave plate for linear polarization measurements .
see manset & bastien ( 1995 ) and manset ( 2000 ) for details about the instrument .
the high voltage supplied to the pockels cell is switched at 62.5 hz to beat down the effects of variable seeing and transparency ( serkowski 1974 ) .
the data were calibrated for instrumental efficiency , instrumental polarization ( due to the telescope s mirrors ) , and origin of position angles using , respectively , a glan - thomson polarizing prism , non - polarized and polarized standard stars .
calibration measurements taken between 1995 and 1999 reveal a very low and stable instrumental polarization under 0.025% @xmath4% , and a relatively stable correction for the position angle between @xmath5 and @xmath6 from one observing run to the other .
the observational errors are calculated from photon statistics , but include uncertainties introduced by the polarimetric efficiency of the instrument , the instrumental polarization , and the calibration for the origin of the position angle .
the relative errors in position angle can be as low as 0.1 , but due to instrumental effects , systematic errors , and the calibration procedure itself , the absolute errors on the position angles are of the order of 1@xmath7 .
details on the data reduction , stability of the instrument , and calibrations are given by manset ( 2000 ) .
we have also obtained data using the polarimeter sterenn at the 2-meter bernard - lyot telescope of the observatoire du pic - du - midi ( opdm ) , france .
this 2-channel polarimeter uses a half - wave plate rotating at 20 hz and a wollaston prism , along with red photo - diodes .
instrumental efficiency is measured with polarized standard stars , and the instrumental polarization , with non - polarized standard stars .
the observational errors are calculated from the fit of the observations taken at a frequency of 20 hz to a sinusoid function .
polarimetric observations are in general a sum of intrinsic and interstellar polarizations .
it is useful to estimate the importance of interstellar polarization in each measurement made .
one can start by looking at the mathewson et al .
( 1978 ) catalog , which contains over 7000 polarimetric observations of bright early - type stars , and compare the level of polarization and position angle of neighboring stars . however , this can be misleading since interstellar polarization depends on the distance ; a target with high polarization could simply be further away and not necessarily intrinsically polarized .
therefore , for each observed pms binary , the catalog was scanned to select at least 20 close stars _ with a similar distance modulus _ ; depending on the stellar density and number of measurements in the catalog , this led to the selection of a region between 1 and 15 in radius around the target .
the stars selected from the catalog are used to find the ratio @xmath8 , and finally , based on an extinction value for our target and assuming this extinction is of interstellar origin only , an estimate of the interstellar polarization for the target is calculated , along with the average interstellar polarization angle .
this angle is calculated with a simple average and with a distance - weighted average of the polarization angles of all the stars selected .
if the alignment is good over all of the region studied , the two values will be similar to within @xmath9 ; if they are very different , it means the alignment is not very good and it is harder to find an average interstellar polarization angle .
if the position angles for the interstellar polarization and for the target are different , it points to an intrinsic origin for at least part of the polarization measured .
intrinsic polarization is also deduced from polarimetric variability .
histograms and maps of the polarization and its position angle are also used to estimate the importance of interstellar polarization .
since a majority of single pms stars are variable polarimetrically ( bastien 1982 ; drissen , bastien , & st - louis 1989 ; mnard & bastien 1992 ) , we expected pms binaries , and mwc 1080 in particular , to also be polarimetrically variable , either periodically or not .
various statistical tests were applied to check the polarimetric variability or stability of pms binaries ; the details of those variability tests and the results for all the binaries of our sample will be presented in future papers .
many pms binary stars show odd observations with polarization levels and/or position angles well below or above the bulk of the data , and mwc 1080 is no exception .
these data were removed before testing for variability .
we believe these observations were due to some eruption - like events or significant modifications in the circumstellar environment ( formation / destruction of condensations , accretion events ) , and not because of an instrumental problem : a close examination of polarization observations taken over 5 years of non - polarized standard stars ( 84 observations ) , polarized standard stars ( 53 observations ) , and 3 stars that were followed for many consecutive hours ( 121 observations ) did not show odd observations like the ones we repeatedly saw for pms binaries in general , and mwc 1080 in particular ( manset 2000 ) .
in addition to stochastic polarimetric variability , which is a general property of single pms stars , pms binaries will also present periodic polarimetric variations caused by the orbital motion , even if in some cases the amplitude may be too small to be detected with currently available instruments or masked by non - periodic or pseudo - periodic variations . in the case of mie scattering , we have also shown ( paper ii ) that dust grains , which are the main producers of the polarization , are less efficient polarizers and produce smaller amplitude variations than electrons ; this is an indication that periodic polarimetric variations could be more difficult to observe in pms binaries than in , for example , hot stars surrounded by electrons , which can easily show variations of a few tenths of a percent ( see for example robert et al . 1990 ; robert et al .
1992 ) . using the known orbital period
, we calculate the orbital phase for each observation of each star and plot @xmath10 , @xmath11 and the stokes parameters @xmath12 and @xmath13 as a function of the orbital phase .
when enough data are available , observations are represented as first and second harmonics of @xmath14 , where @xmath15 is the orbital phase , @xmath16 : + @xmath17 the coefficients of this fit are then used to find the orbital inclination , using the first or second - order fourier coefficients , although it is usually expected that second order variations will dominate ( bme ) : + @xmath18 ^ 2 & = & \frac{(u_1+q_2)^2 + ( u_2-q_1)^2}{(u_2+q_1)^2 + ( u_1-q_2)^2 } \label{eq - io1-p3},\\ \left [ \frac{1-\cos i}{1+\cos i } \right]^4 & = & \frac{(u_3+q_4)^2 + ( u_4-q_3)^2}{(u_4+q_3)^2 + ( u_3-q_4)^2 } \label{eq - io2-p3}.\end{aligned}\ ] ] non - periodic phenomenons such as eruptive events , transient stellar spots , variable accretion , and rearrangements of the circumstellar or circumbinary material can cause pseudo - periodic polarimetric variations that may mask the strictly periodic ones , especially if the observations are taken over many orbital periods , as is the case here .
when periodic variations are detected , we use a phase dispersion method ( stellingwerf 1978 ) and a lomb normalized periodogram algorithm ( press et al .
1997 ) to investigate the significance of this periodicity .
the phase dispersion method ( pdm ) is a least - squared fitting technique suited for non - sinusoidal time variations covered by irregularly spaced observations , and finds the period that produces the least scatter about the mean curve .
the lomb normalized periodogram ( lnp ) method is more powerful than fast fourier transform methods for uneven sampling , but still assumes the curve is sinusoidal , which may not be always appropriate for the polarimetric observations presented here .
herbig ae / be stars are the higher - mass counterparts ( @xmath19m@xmath20 m @xmath21 m@xmath22 ) to the t tauri stars .
these objects are also associated with nebulosity and surrounded by circumstellar material ; for a review , see catala ( 1989 ) . the emission in ir and mm - wave spectral domains
can be explained by the presence of disks , with possibly central holes , and of envelopes ( miroshnichenko et al .
1999 ) .
mwc 1080 is a luminous herbig ae / be object of the be type with @xmath23 l@xmath22 , and a strong wind ( henning et al .
its kinematic distance was determined to be 2.2 kpc ( levreault 1985 ) , and confirmed by photometry of neighboring stars by grankin et al .
there is a third star @xmath24 away , west of mwc 1080 ( leinert et al .
1994 ) , and a poorly collimated herbig - haro flow ( poetzel , mundt , & ray 1992 ) ; the proposed outflow cavity has a position angle of @xmath25 .
mwc 1080 s spectrum is characterized by strong emission lines ( h@xmath26 , h@xmath27 , ) , some with p cygni profiles , but few absorption lines ( herbig 1960 ; poetzel et al .
the strong emission lines imply the presence of a large amount of circumstellar gas in the system .
the emission - line profiles have p cygni shapes which would indicate that the gaseous envelope of mwc1080 is rather nearly spherical or at least not highly flattened , implying a small contribution to the optical polarization .
moreover , from spectroscopic data ( corporon & lagrange 1999 ) , one could assume that the envelope is not very asymmetric .
on the other hand , according to hillenbrand et al .
( 1992 ) , mwc 1080 would belong to the group i herbig ae / be stars , and thus have a geometrically flat , optically thick circumstellar accretion disk , with optically thin inner regions from the stellar surface up to several stellar radii . this classification could change if other sources ( like the other component of this multiple system ) included in the beams or apertures contribute significantly to the observed emission .
harvey , thronson , & gatley ( 1979 ) have reported an extended far - infrared emission associated with this object .
an @xmath28 map by deutsch et al .
( 1995 ) shows extended elliptical emission 4 or 4000 au in diameter . at @xmath29 ,
the emission is also extended ( di francesco et al . 1994 ) .
there is a bipolar molecular outflow ( yoshida et al .
1991 ) .
mwc 1080 is an eclipsing binary with an orbital period of @xmath30d ( grankin et al .
1992 ; shevchenko et al .
1994 ) . since the photometric variations are close to sinusoidal ( grankin et al .
1992 ) , it is thought that the masses are very similar . the asymmetric light curve and different height peaks indicate an eccentric orbit with @xmath31@xmath32 , whereas the moderate amplitude is due to a perceptible inclination of the orbital plane , i.e. , the system is not seen exactly edge - on ( grankin et al .
1992 ) .
hillenbrand et al .
( 1992 ) measured a polarization of 1.79% at 75 in a @xmath33 filter , estimated the interstellar polarization from neighboring stars to be 2.58% at 71 , therefore giving an intrinsic polarization of 0.79% .
concerning this subtraction . ] on the other hand , garrison & anderson ( 1978 ) , using a different method , determined a different interstellar polarization of 2.00% at 168 , giving an intrinsic @xmath33 polarization of 3.84% at 76 .
mwc 1080 was observed at the omm ( canada ) between august 1995 and june 1999 , using a @xmath34 aperture hole and a broad red filter ( filter rg645 centered on @xmath0 , with @xmath35 fwhm ) .
these data are presented in table [ tab - data ] where we give the universal and julian dates , the orbital phase , polarization and position angle along with their uncertainties .
the observations thus include the third star , located at @xmath24 from the spectroscopic binary , or @xmath36au from the binary , assuming a distance of @xmath37kpc ( levreault 1985 ; grankin et al .
since this star is located far away , it can contribute to the observations by adding a constant polarization , but it should not introduce variability unless it experiences some stochastic phenomenon such as flare - like eruptions or stellar spots .
figure [ fig - mwc1080_all ] presents these observations , with the polarization @xmath10 , position angle @xmath11 and the stokes parameters @xmath12 and @xmath13 shown as a function of the orbital phase .
the solid lines are the fourier fits made according to equations [ p3-eq - qfit ] and [ p3-eq - ufit ] . as can be seen on that figure
, mwc 1080 presents a few ( 4 ) observations which stand out from the rest of the data , and which were removed for the variability analysis and the fitting procedure ( data from 1996 , june 2 , july 9 , september 1 and 8) .
the data are also very noisy .
the very deviant observations and the noise are not due to instrumental problems since observations for other stars are stable ( see section [ section - var - tests ] ) and some of the observations for mwc 1080 consist of @xmath19@xmath38 individual measurements which are all consistent with one another .
we have also obtained data at the opdm ( france ) in october 1996 , using a @xmath33 filter and a @xmath39 aperture hole .
these observations are presented in table [ tab - pic - v ] , along with data from the literature and omm .
the @xmath33 observations from omm and opdm are presented in figure [ fig - mwc_v ] .
observations were also obtained at omm on 3 occasions at other wavelengths to produce curves of the linear polarization as a function of wavelength ( see table [ tab - p_lambda ] ) .
an inspection of the mathewson et al .
( 1978 ) catalog shows that mwc 1080 is in a region where the position angles are well aligned at @xmath40@xmath41 ( simple and weighted averages of the position angles for 51 stars found within 4 of mwc 1080 and having similar distance modulus ) , with a wide distribution of polarization values that go up to 5% ( see figure [ fig - pol_is ] ) .
this indicates that the interstellar polarization is strong in this region , and we estimate it to be , based on polarization for neighboring stars , @xmath42% at 73 in the @xmath33 band , similar to the estimation by hillenbrand et al .
( 1992 ) of 2.58% at 71 in the @xmath33 band , but perpendicular to the garrison & anderson ( 1978 ) value of 2.00% at 168 .
we believe , based on hillenbrand et al.s ( 1992 ) and our analysis of the mathewson et al .
( 1978 ) catalog , that the interstellar polarization has a position angle of @xmath43 and not @xmath44 .
using our estimate of the is polarization , 2.21% at 73 in the @xmath33 band , and serkowski s law for interstellar polarization ( serkowski , mathewson , & ford 1975 ) , we find an interstellar polarization value of 1.95 % at @xmath45 at @xmath0 .
we have assumed that @xmath46 for serkowski s law ( a typical value , and also a mean for 3 stars found near mwc 1080 in the mathewson et al .
( 1978 ) catalog , and a value compatible with our wavelength - dependent data ) and that the position angle of the interstellar polarization is constant as a function of wavelength . from our data , mwc 1080 s average polarization is 1.60% at 81.6at @xmath0 ( @xmath47 ) , and 1.96% at 76.4 at @xmath33 ( @xmath48 , data from opdm and omm only ) .
nine stars are found within 1 of mwc 1080 .
two , at about half a degree , show polarization levels of 1.36% and 1.54% , with position angles of 62 and 63 , which indicates that a significant part of mwc 1080 s polarization is of interstellar origin . a proper subtraction of the interstellar polarization from the observed polarization ( using the stokes parameters @xmath49 and @xmath50 and not a usual vectorial subtraction as was apparently done in the hillenbrand et al .
( 1992 ) paper , in particular in their table 2 ) gives estimates of the average intrinsic polarization for mwc 1080 : 0.6% at 139 at @xmath0 , and 0.35% at 142 in the @xmath33 band .
a proper subtraction taking the hillenbrand et al .
( 1992 ) @xmath33 data would give @xmath51 at 152 ; garrison & anderson ( 1978 ) report a much higher and perpendicular polarization , 3.84% at 76 but we believe this is due to an inappropriate evaluation of the interstellar polarization .
the position angles @xmath52 trace the axis of symmetry of the system and are at about 90 with respect to the position angle ( @xmath25 ) of the outflow cavity ( poetzel , mundt , & ray 1992 ) , which is what is usually observed for pms stars ( bastien 1988 ) .
we then confirm the poetzel et al .
( 1992 ) scenario where the elongated structure they saw in the continuum at @xmath53 is made of the innermost and brightest part of a reflection nebula , elongated in the outflow direction , and perpendicular to the polarization vector .
the significant intrinsic polarization indicates that the circumstellar or circumbinary material around this binary is arranged in an asymmetric geometry , for example , in a disk rather than a spherical shell .
two other facts point to an intrinsic polarization component in the observed polarization for mwc 1080 .
first , mwc 1080 shows clear periodic variations ( see below ) that can not be of interstellar origin .
second , the position angle of the polarization varies as a function of wavelength ( see below ) .
observations of polarized standard stars ( whose polarization is of interstellar origin ) by schmidt , elston , & lupie ( 1992 ) show that the maximum observed rotation of the position angle for the wavelength range covered by our observations ( 4350@xmath54 ) is @xmath55 , much less than the @xmath56 rotation seen for mwc 1080 .
as discussed by dolan & tapia ( dolan & tapia 1986 and references cited ) , a wavelength - dependent position angle could in theory be attributed to multiple interstellar clouds containing different grain sizes magnetically aligned in different directions , or to an interstellar cloud where there is a continuous rotation of the grains orientation , but , once again , the variability argument points to the presence of intrinsic polarization . already from figure [ fig - mwc1080_all ] , it is obvious that mwc 1080 s polarization is variable .
the detailed results of the variability tests for the @xmath0 wavelength and the @xmath33 filter will be presented in a future paper , along with analysis for our other binaries .
the analysis shows that mwc 1080 was clearly variable polarimetrically between august 1995 and june 1999 , in both filters , with indications of variability on time scales of days and months .
the average polarization at @xmath0 for the period 19951999 is 1.6% at 82 , whereas unpublished data taken at the omm on jan .
15 1988 in the same filter and aperture hole show a significantly higher polarization value , @xmath57% , but with a similar position angle , 81.7 .
this points to long - term ( years ) variability .
data for the @xmath33 filter show that there might be a significant difference between observations taken before and after october 1973 , as data from 1973 ( vrba et al .
1975 ) show a higher polarization and ( slightly ) different position angle .
after 1973 , the polarization and its angle seem to have stabilized , although there is still evidence of variability .
in particular , the four observations taken at opdm on 4 consecutive nights clearly show variability . once again
, this indicates that variability is present with different time scales . on some nights ( 6 ) , it was possible to get 3 observations within about 5 hours . on many occasions ,
a set of 3 points shows a systematic increase or decrease in polarization and/or position angle , usually by a significant amount , showing that the variations are not random . on other occasions , the position angle stayed constant .
this indicates that variability is present on time scales of hours , as well as days , months , and years .
the variations in position angle are confined to within @xmath58 of the mean , which indicates that polarization variations are probably not due to a drastic change in the geometry of the circumstellar material , but rather to modifications in the density .
figure [ fig - mwc1080_all ] shows all the @xmath0 observations , and figure [ fig - mwc1080_bin20 ] shows binned data , where the orbital phase has been divided in equal bins and the polarization data weight averaged ; the error bars in figure [ fig - mwc1080_bin20 ] are simple averages of the error bars of the data in each bin .
clear periodic variations with amplitudes in @xmath12 and @xmath13 @xmath59% are seen , especially in polarization angle , but are not fitted well by the simple @xmath60 and @xmath61 sinusoidal curves ; the coefficients for this fit are ( equations [ p3-eq - qfit ] and [ p3-eq - ufit ] ) : adding @xmath62 and @xmath63 harmonics does a better job ( see figure [ fig - mwc1080_fit4 ] ) , but the fit is still not adequate ; the coefficients for that fit are : the level of the the polarimetric variations could point to a circumstellar disk rather than a circumbinary disk , as we have shown in paper ii that the former configuration is more favorable to high amplitude variations .
however , if we assume both stars in mwc 1080 have masses of @xmath64m@xmath22 , the orbital period yields a separation of @xmath65au between each star ; if the radii are @xmath66r@xmath22 , or about @xmath67au , the photospheres are separated by only @xmath68au , which does not leave much space for dusty circumstellar material ( which might then just evaporate due to the proximity of the photospheres ) . in the observed polarimetric curves ,
there are minima of polarization and maxima in position angle near phases 0.0 and 0.5 .
when the interstellar polarization , which is almost perpendicular to the observed one , is removed from the observed polarization , the intrinsic polarization now shows a maxima near phases 0.0 and 0.5 , close to the position of the primary and secondary eclipses ( 0.0 and 0.55 ( shevchenko et al .
both @xmath0 and @xmath33 observations show a sudden increase in intrinsic polarization around phase 0.5 ( see figures [ fig - mwc_v ] and [ fig - mwc1080_bin20 ] ) .
since this change is seen in both filters , it must be related to the eclipses .
an increase in polarization during an eclipse is to be expected since shielding the unpolarized light from the star will increase the amount of the scattered ( and polarized ) light .
increase of polarization at phase 0.0 had already been detected in the @xmath13 , @xmath69 , and @xmath33 bands by shevchenko et al .
( 1994 ) .
a prediction of our paper i was that as the orbital eccentricity increases , the periodic variations start from pure double - periodic to include stronger and stronger single - periodic variations .
the eccentricity for mwc 1080 is not known but is estimated to be between 0.2 and 0.5 ( grankin et al .
1992 ) ; the strength of double - periodic variations , which have amplitudes between 1.4 and 2.5 times those of the single - periodic one , would favor an eccentricity nearer to 0.2 than to 0.5 , but the polarimetric curves can not be formally inverted to find the orbital eccentricity .
in addition , non - periodic variations introduce noise that contributes , sometimes substantially , to both harmonics , so that such a straightforward association between eccentricity and presence of @xmath60 variations can not be made .
we have shown in papers i and ii that factors other than eccentricity ( time - varying optical depth , non equal luminosity stars , asymmetric geometries , scattering on dust grains ) can also introduce such single - periodic variations .
the pdm finds periods of 1.1 and @xmath70d when we look at a subset of the data taken over a few months only , and the lnp shows a peak at @xmath71d which becomes significant ( more than 97% chance that the data do not come from random gaussian noise ) when we look at the binned data .
the periods of @xmath71d and @xmath70d could be related to the orbital period of @xmath72d . on three occasions , in 1996 , 1997 , and 1998 , data were obtained in order to get @xmath73 , the polarization as a function of wavelength .
data are presented in table [ tab - p_lambda ] and figure [ fig - p_lambda ] .
the polarization and its position angle are clearly variable as a function of wavelength and time .
the observed position angle generally decreases towards shorter wavelengths , by up to 10 between @xmath74 and @xmath54 .
changes of up to 10 are also seen from year to year .
variations like those have already been observed for t tauri stars ( bastien 1981 ) .
the change of polarization angle as a function of wavelength is mostly an intrinsic variation , since the interstellar polarization shows minute variations as a function of wavelength ( once again , see schmidt , elston , & lupie 1992 ) .
also , a study of the variability of polarized standard stars ( bastien et al . 1988 ) , where the variability might due to is polarization variability , shows variations @xmath75 , also below those seen for mwc 1080 .
we therefore believe that is polarization variability would not be able to explain the change of position angle as a function of wavelength for mwc 1080 . as data taken during the same night at the same wavelength show ,
significant variations might have occurred during the time the @xmath73 was measured .
nonetheless , the 3 curves seem to be relatively consistent from one data set to the other .
the polarization has a general bell shape usually associated with interstellar polarization and the serkowski law ( serkowski , mathewson , & ford 1975 ) .
in particular , this interstellar curve has a peak at @xmath76 , near the peak of mwc 1080 s curve , which indicates , as already stated , that mwc 1080 s polarization is in great part of interstellar origin .
using our estimate of the is polarization for the @xmath33 band and serkowski s law , we can compute the is polarization for the observed wavelengths , and subtract it to find the intrinsic @xmath73 ( see figure [ fig - mwc_instrinsic ] ) .
removal of the is polarization moves the peak of polarization to a redder wavelength around @xmath77 . using the numerical code presented in papers i and
ii , we have computed a few @xmath73 curves for different grain composition and sizes . even though the number of free parameters is too high to uniquely determine the composition and size of the grains from the observed @xmath73 curve , these curves
can nonetheless give some insight into the environment for mwc 1080 .
figure [ fig - p_lambda_as ] presents 3 @xmath73 curves for astronomical silicate grains of radii @xmath78 , @xmath79 , and @xmath80 .
it is immediately seen that the size of the grains has a profound impact on the morphology of that curve . from our simulations ,
the circumbinary disk configuration does not produce variations of the polarization angle as a function of wavelength , but the circumstellar disk does , for certain orbital phases .
the variations are small , 34 , but the dependence can be either increasing or decreasing toward the blue , depending on the orbital phase . since mwc 1080 does show a variable polarization angle , it points to a circumstellar disk , or at least , to the presence of matter close to the stars .
qualitatively , then , the polarization for mwc 1080 could be reproduced by astronomical silicate grains of small sizes , between @xmath78 and @xmath79 , located close to the stars , in an asymmetric configuration .
photometric observations of the irregular variable mwc 1080 show periodic variations with normal light variations of 0.16 magnitudes in @xmath33 , along with irregular variability with @xmath81 magnitudes , which introduces scatter in the periodic variations ( grankin et al .
an inspection of a catalog of @xmath82 photometry of t tauri stars ( herbst et al .
1994 ) with its on - line database containing @xmath83 observations for this binary also reveals that the average magnitude for mwc 1080 can change from one epoch to another .
peak - to - peak variations are @xmath84@xmath85 magnitudes in @xmath33 , whereas the scatter is of the same order .
this stochastic photometric variability is probably related to the stochastic polarimetric variability observed here .
we have tried to cross - correlate photometric observations taken from that database with our polarimetric observations , and find data obtained quasi - simultaneously . due to the physical locations of the observatories , the closest observations are within @xmath86@xmath32 days of each other , which is a significant portion of the orbital period . nonetheless , we found about half a dozen quasi - simultaneous observations , which do not show any clear correlation between brightness of the star and its intrinsic polarization .
one of the goals of these observations is to find the orbital inclination for this binary system .
since it is an eclipsing binary , the orbital inclination is known to be close to 90 , but the light curve also suggests some small departure from an edge - on configuration ( grankin et al .
1992 ) . since the orbit might be significantly eccentric ( @xmath87 , grankin et al .
1992 ) , we should consider the results of the bme formalism using both orders ( equations [ eq - io1-p3 ] and [ eq - io2-p3 ] ) , which are @xmath88 and @xmath89 .
if the orbital eccentricity is @xmath90@xmath85 , both orders should give about the same inclination , which is not the case .
if the orbital eccentricity is higher than @xmath91 , only the first - order results should be used ( @xmath88 ) , but this result is not compatible with an eclipsing system , although the separation between the two components is small . in paper
i , we showed that the amplitude of the polarimetric variations and the stochastic noise ( or , scatter of the observations about the mean polarimetric variations ) must be considered first in order to see if the orbital inclination found from equations [ eq - io1-p3 ] and [ eq - io2-p3 ] is meaningful or not . according to our noise criteria ( found after adding different levels of noise in the polarimetric curves produced by our numerical simulations
; see paper i for more details ) , the noise present in mwc 1080 s data is too high ( the scatter is @xmath92% in polarization , or 50% of the amplitude of the variations ) to allow the bme formalism to be used . even applying the same analysis to the binned data , less numerous but showing much less scatter , gives the same conclusion .
the data have to be of exquisite quality ( high number of observations , clear variations , not much stochastic noise ) in order for the inclination found with the bme formalism to be meaningful . in the case of hot stars ,
the variations are usually of higher amplitude ( by a factor of @xmath93 ) with less scatter from intrinsic non - periodic variations , so that in those cases , the noise criteria indicates the bme results are meaningful .
we have presented polarimetric observations of a young massive short - period binary star of the herbig ae / be type , mwc 1080 .
the mean polarization at @xmath0 is 1.60% at 81.6 ( or 0.63% at 139 if an estimate of the interstellar polarization is subtracted ) and 1.96% at 76.4 in the @xmath33 band .
the significant intrinsic polarization points to an asymmetric geometry of the circumstellar or circumbinary environment .
the 139 intrinsic position angle traces the axis of symmetry of the system and is perpendicular to the position angle of the outflow cavity ( @xmath94 ) .
although the interstellar polarization is very strong in this region , mwc 1080 s polarization can not be of interstellar origin only , for 3 reasons .
first , the estimated interstellar polarization for this region is different than the observed one for mwc 1080 .
second , the significant variability of the polarization as a function of time ( especially on time scales of hours ) can not be explained by modifications in the interstellar medium .
third , the dependence of the position angle on the wavelength along with its variability can not be attributed to the interstellar medium .
this polarization is clearly variable , at all wavelengths , and on time scales of hours , days , months , and years ; this is not a surprise and was expected since many pms stars , of the herbig ae / be or t tauri type , show such variations .
in addition to that stochastic variability , there are periodic variations due to the orbital motion of the stars in their dusty environment . although stochastic ( non - periodic ) polarimetric variations introduce noise and scatter around the mean variations , binning the numerous observations reveals clear periodic polarimetric variations , which follow the known orbital period . the periodic polarimetric variations of this binary star are the first phased - locked ones detected for a pms binary star that the authors are aware of .
the amplitude of the variations point to a circumstellar disk rather than a circumbinary one , since , according to our numerical simulations , the second configuration does not produce significant variations .
however , the stars are probably so close to one another that there is not much room left for circumstellar material .
the variations are not simply double - periodic ( as produced by a simple model of 2 equal mass stars in a circular orbit , at the center of an axisymmetric circumbinary envelope made of electrons ) , but include single - periodic and higher order variations .
the presence of single - periodic variations could be due to non equal mass stars ( although photometric observations point to equal mass or similar mass stars ) , the presence of dust grains , an asymmetric configuration of the circumstellar or circumbinary material , or the eccentricity of the orbit ( which is thought to be between 0.2 and 0.5 ) .
the available data are for the moment insufficient to say what is the predominant cause of those single - periodic variations , or to find the geometry or orbital eccentricity of this system .
mwc 1080 is an eclipsing binary with primary and secondary eclipses occurring at phases 0.0 and 0.55 .
the signatures of the eclipses are seen in the polarimetric observations .
the primary eclipse coincides with maxima in intrinsic polarization and position angle in the @xmath0 data .
in addition , both @xmath0 and @xmath33 observations show a sudden increase in polarization around phase 0.5 where the secondary eclipse is located .
most of the observations have been obtained at a wavelength of @xmath0 , but we have obtained additional @xmath33 band observations , and also observations as a function of wavelength on 3 occasions .
the @xmath73 curve has a bell - shape morphology which changes on a time scale of a year , with a maximum polarization at about @xmath95 , typical of the interstellar polarization ; when the estimated interstellar polarization is removed , the peak in the @xmath73 curve is displaced towards @xmath96 .
the position angle also depends on the wavelength , with a rotation of @xmath97 over @xmath98 , and on time . both mwc 1080 s polarization and position angle are thus dependent on time and wavelength .
the intrinsic polarization and intrinsic @xmath73 curves are consistent with the presence of dusty material in the circumstellar environment .
although it is not possible to uniquely determine the composition and size of the dust grains , our observations would be compatible with astronomical silicate grains with sizes in the range @xmath99@xmath100 .
photometry for this irregular variable star shows stochastic variations that introduce significant scatter in the photometric light curve , a characteristic also seen in our polarimetric data .
no clear correlation between brightness and polarization was found .
the stochastic polarimetric variations mentioned before introduce so much scatter that the orbital inclination obtained with the bme formalism is unfortunately not meaningful .
this example shows that polarimetric observations have to be of exquisite quality ( high number of observations , clear variations , not much stochastic noise ) in order to find the orbital inclination .
n. m. would like to thank the conseil de recherche en sciences naturelles et gnie of canada , the fonds pour la formation de chercheurs et laide la recherche of the province of qubec , the facult des etudes suprieures and the dpartement de physique of universit de montral for scholarships , and p. b. for financial support .
we would like to thank the conseil de recherche en sciences naturelles et gnie of canada for supporting this research .
the authors gratefully acknowledge financial support from universit joseph fourier , grenoble ( 1994 bqr grant b 644 r1 ) , the laboratoire dastrophysique de lobservatoire de grenoble , and the observatoire midi - pyrnes for upgrade and maintenance of the sterenn polarimeter .
bastien , p. 1981
, , 94 , 294 bastien , p. 1982
, , 48 , 153 bastien , p. 1988
, in polarized radiation of circumstellar origin , ed
. g. v. coyne et al .
( vatican city : vatican observatory ) , 541 bastien , p. 1996 , in asp conf .
97 , polarimetry in the interstellar medium , ed .
w. g. roberge & d. c. b. whittet ( san francisco : asp ) , 297 bastien , p. , drissen , l. , mnard , f. , moffat , a. f. j. , robert , c. , & st - louis , n. 1988 , , 95 , 900 brooks , a. , clarke , d. , & mcgale , p. a. 1994 , vistas in astronomy , 38 , 377 brown , j. c. , mclean , i. s. , & emslie , a. g. 1978 , , 68 , 415 catala , c. 1989 , in eso conf .
33 , low mass star formation and pre - main - sequence evolution , ed .
b. reipurth ( garching : eso ) , 471 clarke , d. , & naghizadeh - khouei , j. 1994 , , 108 , 687 corporon , p. & lagrange a .- m . 1999 ,
suppl . , 136 , 429 deutsch , l. k. , hora , j. l. , butner , h. m. , hoffman , w. f. , & fazio , g. g. 1995 , , 224 , 89 di francesco , j. , evans , n. j. , harvey , p. m. , mundy , l. g. , & butner , h. m. 1994 , , 432 , 710 dolan , j. f. , & tapia , s. 1986 , , 98 , 792 drissen , l. , bastien , p. , & st - louis , n. 1989 , , 97 , 814 garrison jr . , l. m. , & anderson , c. m. 1978 , , 221 , 601 grankin , k. n. , shevchenko , v. s. , chernyshev , a. v. , ibragimov , m. a. , kondratiev , w. b. , melnikov , s. yu . , & yabukov , s. d. 1992 , iau information bull . on variable stars 3747 harvey , p. m. , thronson , h. , jr . ,
& gatley , i. 1979 , , 231 , 115 heiles , c. 2000 , , 119 , 923 henning , th . ,
burkert , a. , launhardt , r. , leinert , ch . , & stecklum , b. 1998 , , 336 , 565 herbig , g. h. 1960 , , 43 , 337 herbst , w. , herbst , d. k. , grossman , e. j. , & weinstein , d. 1994 , , 108 , 1906 hillenbrand , l. a. , strom , s. e. , vrba , f. j. , & keene , j. 1992 , , 397 , 613 leinert , ch . ,
richichi , a. , weitzel , n. , & haas , m. 1994 , in asp conf .
62 , the nature and evolutionary status of herbig ae / be stars , ed .
p. s. th , m. r. perez , & p. j. van den heuvel ( san francisco : asp ) , 155 levreault , r. m. 1985 , ph.d .
thesis , university of texas at austin manset , n. 2000 , ph.d .
thesis , universit de montral manset , n. , & bastien , p. 1995
, , 107 , 483 manset , n. , & bastien , p. 2000 , , 120 , 413 ( paper i ) manset , n. , & bastien , p. 2001
, , in press ( astro - ph/0108124 ) ( paper ii ) mathewson , d. s. , ford , v. i. , klare , g. , neckel , th . , & krautter , j. 1978 , bicds , 14 , 115 mnard , f. , & bastien , p. 1992
, , 103 , 564 miroshnichenko , a. , ivezi , . , vinkovi , d. , & elitzur , m. 1999 , , 520 , l115 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 1997
, numerical recipes in c , the art of scientific computing , 2d edition ( cambridge : cambridge university press ) poetzel , r. , mundt , r. , & ray , t. p. 1992
, , 262 , 229 robert , c. , moffat , a. f. j. , bastien , p. , st - louis , n. , & drissen , l. 1990 , , 359 , 211 robert , c. , et al .
1992 , , 397 , 277 rudy , r. j. , & kemp , j. c. 1978 , , 221 , 200 schmidt , g. d. , elston , r. , & lupie , o. l. 1992 , , 104 , 1563 serkowski , k. 1958 , acta astronomica , 8 , 135 serkowski , k. 1974 , in iau coll .
23 , planets , stars , and nebulae studied with photopolarimetry , ed .
t. gehrels ( tucson : university of arizona press ) , 135 serkowski , k. , mathewson , d. s. , ford , v. l. 1975 , , 196 , 261 shevchenko , v. s. , grankin , k. n. , ibragimov , m. a. , kondratiev , v. b. , melnikov , s. yu . ,
petrov , p. p. , shcherbakov , v. a. , & vitrichenko , e. a. 1994 , in asp conf .
ser . , 62 ,
the nature and evolutionary status of herbig ae / be stars ( san francisco : asp ) , 43 stellingwerf , r. f. 1978 , , 224 , 953 topping , j. 1972 , errors of observation and their treatment ( london : chapman and hall ) vrba , f. j. 1975 , , 195 , 101 yoshida , s. , kogure , t. , nakano , m. , tatematsu , k. , & wiramihardja , s. 1991 , , 43 , 363 lcccccc 1995 aug 17 & 49946.824 & 0.138 & 1.546 & 0.029 & 80.7 & 1.0 + 1995 aug 19 & 49948.809 & 0.826 & 1.641 & 0.021 & 81.6 & 1.0 + 1995 aug 26 & 49955.683 & 0.207 & 1.687 & 0.029 & 81.9 & 1.0 + 1995 aug 28 & 49957.646 & 0.887 & 1.577 & 0.026 & 82.1 & 1.0 + 1995 aug 29 & 49958.617 & 0.223 & 1.612 & 0.026 & 81.5 & 1.0 + 1995 aug 31 & 49960.588 & 0.906 & 1.557 & 0.024 & 82.2 & 1.0 + 1995 sep 2 & 49962.833 & 0.683 & 1.684 & 0.020 & 80.4 & 1.0 + 1995 sep 3 & 49963.567 & 0.938 & 1.564 & 0.024 & 83.2 & 1.0 + 1995 sep 3 & 49963.774 & 0.010 & 1.616 & 0.024 & 82.1 & 1.0 + 1995 sep 4 & 49964.616 & 0.301 & 1.570 & 0.019 & 80.9 & 1.0 + 1996 jan 1 & 50084.490 & 0.824 & 1.775 & 0.031 & 84.4 & 1.0 + 1996 jan 3 & 50085.588 & 0.205 & 1.706 & 0.028 & 82.4 & 1.0 + 1996 jan 4 & 50087.483 & 0.861 & 1.643 & 0.028 & 81.9 & 1.0 + 1996 jan 7 & 50089.488 & 0.555 & 1.698 & 0.032 & 85.6 & 1.0 + 1996 jun 1 & 50235.768 & 0.225 & 1.630 & 0.024 & 81.8 & 1.0 + 1996 jun 2 & 50236.760 & 0.569 & 2.205 & 0.045 & 75.0 & 1.0 + 1996 jun 2 & 50236.787 & 0.578 & 1.590 & 0.031 & 81.0 & 1.0 + 1996 jun 12 & 50246.769 & 0.036 & 1.480 & 0.028 & 84.9 & 1.0 + 1996 jul 9 & 50273.682 & 0.358 & 2.108 & 0.041 & 71.4 & 1.0 + 1996 aug 24 & 50319.730 & 0.309 & 1.624 & 0.022 & 81.2 & 1.0 + 1996 aug 25 & 50320.592 & 0.607 & 1.549 & 0.025 & 82.0 & 1.0 + 1996 aug 25 & 50320.678 & 0.637 & 1.642 & 0.024 & 81.6 & 1.0 + 1996 aug 25 & 50320.808 & 0.682 & 1.721 & 0.028 & 81.2 & 1.0 + 1996 sep 1 & 50327.630 & 0.045 & 1.557 & 0.033 & 80.6 & 1.0 + 1996 sep 1 & 50327.742 & 0.084 & 1.699 & 0.029 & 80.3 & 1.0 + 1996 sep 1 & 50327.853 & 0.123 & 1.931 & 0.032 & 82.4 & 1.0 + 1996 sep 4 & 50330.614 & 0.079 & 1.626 & 0.027 & 81.2 & 1.0 + 1996 sep 4 & 50330.720 & 0.116 & 1.598 & 0.027 & 80.0 & 1.0 + 1996 sep 4 & 50330.846 & 0.159 & 1.644 & 0.029 & 80.8 & 1.0 + 1996 sep 6 & 50332.583 & 0.761 & 1.610 & 0.027 & 79.9 & 1.0 + 1996 sep 7 & 50333.563 & 0.100 & 1.565 & 0.028 & 83.8 & 1.0 + 1996 sep 7 & 50333.568 & 0.102 & 1.606 & 0.033 & 84.0 & 1.0 + 1996 sep 7 & 50333.685 & 0.143 & 1.536 & 0.027 & 84.0 & 1.0
+ 1996 sep 7 & 50333.820 & 0.189 & 1.597 & 0.029 & 83.9 & 1.0 + 1996 sep 8 & 50334.587 & 0.455 & 1.354 & 0.030 & 84.8 & 1.0 + 1996 sep 8 & 50334.715 & 0.499 & 1.532 & 0.025 & 73.6 & 1.0 + 1996 sep 8 & 50334.721 & 0.501 & 1.575 & 0.029 & 83.5 & 1.0 + 1996 oct 12 & 50368.637 & 0.250 & 1.469 & 0.021 & 83.4 & 1.0 + 1996 oct 16 & 50372.532 & 0.599 & 1.597 & 0.022 & 81.8 & 1.0 + 1996 dec 22 & 50439.523 & 0.804 & 1.727 & 0.027 & 81.2 & 1.0 + 1997 jun 3 & 50602.791 & 0.358 & 1.374 & 0.047 & 81.5 & 1.0 + 1997 jun 4 & 50603.799 & 0.707 & 1.581 & 0.044 & 80.9 & 1.0 + 1997 jun 5 & 50604.772 & 0.044 & 1.514 & 0.033 & 81.5 & 1.0 + 1997 jun 6 & 50605.784 & 0.395 & 1.452 & 0.027 & 80.5 & 1.0 + 1997 jun 7 & 50606.786 & 0.742 & 1.465 & 0.035 & 79.0 & 1.0 + 1997 jun 9 & 50608.774 & 0.431 & 1.434 & 0.028 & 79.7 & 1.0 + 1997 jun 15 & 50614.773 & 0.509 & 1.460 & 0.030 & 79.7 & 1.0 + 1997 jun 16 & 50615.756 & 0.849 & 1.679 & 0.036 & 80.0 & 1.0 + 1997 jul 6 & 50635.753 & 0.776 & 1.525 & 0.025 & 79.3 & 1.0 + 1997 jul 8 & 50637.788 & 0.481 & 1.330 & 0.028 & 83.3 & 1.0 + 1997 jul 10 & 50639.777 & 0.170 & 1.693 & 0.037 & 82.2 & 1.0 + 1997 jul 11 & 50640.674 & 0.480 & 1.693 & 0.027 & 81.1 & 1.0 + 1997 aug 25 & 50685.723 & 0.085 & 1.605 & 0.017 & 83.9 & 1.0 + 1997 sep 6 & 50697.623 & 0.207 & 1.816 & 0.027 & 78.2 & 1.0 + 1997 sep 8 & 50699.823 & 0.969 & 1.536 & 0.033 & 81.6 & 1.0 + 1997 sep 9 & 50700.575 & 0.229 & 1.624 & 0.024 & 81.1 & 1.0 + 1997 sep 9 & 50700.707 & 0.275 & 1.673 & 0.024 & 80.5 & 1.0 + 1997 sep 9 & 50700.820 & 0.314 & 1.553 & 0.024 & 79.7 & 1.0 + 1997 oct 31 & 50752.532 & 0.227 & 1.630 & 0.030 & 81.4 & 1.0 + 1997 nov 4 & 50756.600 & 0.636 & 1.723 & 0.030 & 81.7 & 1.0 + 1998 jun 12 & 50976.666 & 0.864 & 1.505 & 0.108 & 78.0 & 2.1 + 1998 jul 15 & 51009.749 & 0.324 & 1.864 & 0.056 & 77.1 & 1.0 + 1998 aug 28 & 51053.765 & 0.571 & 1.662 & 0.029 & 81.9 & 1.0 + 1999 jun 11 & 51340.795 & 0.995 & 1.532 & 0.046 & 84.1 & 1.0 + 1999 jun 12 & 51341.770 & 0.332 & 1.734 & 0.046 & 81.4 & 1.0 + 1999 jun 13 & 51342.758 & 0.675 & 1.879 & 0.043 & 82.0 & 1.0 lcccccccc oct 73 & 10 & & & 2.54 & 0.3 & 68.2 & & 1 + jan 76 & 13 & & & 1.87 & 0.15 & 74 & & 2 + & 10 & & & 1.79 & & 75.0 & & 3 + 1996 oct 9 & 10&50366.464 & 0.497 & 1.882 & 0.038 & 77.0 & 1.0 & 4 + 1996 oct 10 & 10&50367.373 & 0.812 & 1.985 & 0.043 & 75.6 & 1.0 & 4 + 1996 oct 11 & 10&50368.402 & 0.168 & 2.110 & 0.062 & 78.5 & 1.0 & 4 + 1996 oct 12 & 10&50369.470 & 0.538 & 2.103 & 0.057 & 78.9 & 1.0 & 4 + 1996 oct 16 & @xmath34 & 50372.589 & 0.619 & 2.044 & 0.049 & 75.1 & 1.0 & 5 + 1997 jul 11 & @xmath34 & 50640.738 & 0.503 & 1.713 & 0.047 & 75.9 & 1.0 & 5 + 1998 jul 3 & @xmath34 & 50997.649 & 0.133 & 2.119 & 0.068 & 74.0 & 1.0 & 5 lcccccccc 1996 oct 16 & 50372.626 & 0.632 & 4350(990 ) & 1.910&0.225&70.9&3.4 + & 50372.589 & 0.619 & 5480(1060 ) & 2.044&0.049&75.1&1.0 + & 50372.532 & 0.600 & 7660(2410 ) & 1.597&0.022&81.8&1.0 + & 50372.561 & 0.609 & 8580(630 ) & 1.441&0.050&75.4&1.0 + 1997 jul 11 & 50640.711 & 0.494 & 4350(990 ) & 1.632&0.116&72.4&2.0 + & 50640.738 & 0.503 & 5480(1060 ) & 1.713&0.047&75.9&1.0 + & 50640.787 & 0.520 & 5900(800 ) & 1.786&0.053&79.1&1.0 + & 50640.808 & 0.527 & 7610(915 ) & 1.439&0.053&83.4&1.0 + & 50640.762 & 0.511 & 8580(630 ) & 1.400&0.049&85.6&1.0 + 1998 jul 3 & 50997.712 & 0.155 & 4350(990 ) & 1.945&0.157&76.7&2.3 + & 50997.649 & 0.133 & 5480(1060 ) & 2.119&0.068&74.0&1.0 + & 50997.763 & 0.173 & 8580(630 ) & 1.533&0.072&81.5&1.3 + & 50997.793 & 0.183 & 7610(915 ) & 1.558&0.046&79.5&1.0 + | we present polarimetric observations of a massive pre - main sequence short - period binary star of the herbig ae / be type , mwc 1080 .
the mean polarization at @xmath0 is 1.60% at 81.6 , or 0.6% at 139 if an estimate of the interstellar polarization is subtracted .
the intrinsic polarization points to an asymmetric geometry of the circumstellar or circumbinary environment while the 139 intrinsic position angle traces the axis of symmetry of the system and is perpendicular to the position angle of the outflow cavity .
the polarization and its position angle are clearly variable , at all wavelengths , and on time scales of hours , days , months , and years .
stochastic variability is accompanied by periodic variations caused by the orbital motion of the stars in their dusty environment .
these periodic polarimetric variations are the first phased - locked ones detected for a pre - main sequence binary .
the variations are not simply double - periodic ( seen twice per orbit ) but include single - periodic ( seen once per orbit ) and higher - order variations
. the presence of single - periodic variations could be due to non equal mass stars , the presence of dust grains , an asymmetric configuration of the circumstellar or circumbinary material , or the eccentricity of the orbit .
mwc 1080 is an eclipsing binary with primary and secondary eclipses occurring at phases 0.0 and 0.55 .
the signatures of the eclipses are seen in the polarimetric observations . |
[ s1 ] the tricomi operator @xmath1 one of the simplest examples of a partial differential operator of _ mixed type _ , is : ( i ) _ elliptic _ in the upper half plane ( @xmath2 ) ; _ parabolic _ along the @xmath3-axis ( @xmath4 ) ; and ( iii ) _ hyperbolic _ in the lower half plane ( @xmath5 ) .
our aim is to obtain explicit solutions in the sense of distributions or generalized functions of the equation @xmath6 where @xmath7 is the dirac function at @xmath8 an arbitrary point in the plane . a solution @xmath9 of ( [ eq1c ] ) is said to be a _ fundamental solution relative to the point @xmath8 . _ in a previous paper @xcite we considered the case when @xmath10 and proved the existence of two remarkable fundamental solutions that clearly reflect the fact that the operator changes type across the @xmath3-axis . in this paper
we study the general case when @xmath11 and compare our results to those of @xcite . for sake of completeness and in order to make the reading of this paper independent of that of @xcite
, we briefly review the contents of that paper .
it is known that the equation @xmath12 defines the two characteristics for the tricomi operator that emanate from the origin .
these characteristics , which are tangent to the @xmath13-axis at the origin , divide the plane in two disjoint regions @xmath14 and @xmath15 ( figure 1 ) defined as follows : @xmath16 the region `` outside '' the characteristics , and @xmath17 the region `` inside '' the characteristics .
figure 1 the first fundamental solution is defined by @xmath18 with @xmath19 and the second one is defined by @xmath20 with @xmath21 in the expressions of both @xmath22 and @xmath23 the constant @xmath24 is the value of the hypergeometric function @xmath0 at @xmath25 we remark that in @xcite the constants @xmath22 and @xmath26 were respectively denoted by @xmath27 at that time , we were unaware of the role played by the hypergeometric function @xmath0 in the study of fundamental solutions for the tricomi operator .
we also observe that according to formula ( [ eq1b ] ) , a perturbation at the origin spreads out to the entire region @xmath28 inside the characteristics . on the other hand , according to formula ( [ eq1a ] ) , the same perturbation spreads out to the whole elliptic region ( @xmath2 ) and also to the hyperbolic region ( @xmath5 ) outside the characteristics . since the tricomi operator is invariant under translations along the @xmath3-axis , the same phenomena take place for fundamental solutions that correspond to the dirac measure @xmath29 concentrated at an arbitrary point @xmath30 on the @xmath3-axis
. it is well know that the tricomi operator describes the transition from subsonic flow ( elliptic region ) to supersonic flow ( hyperbolic region ) . in the extensive literature on the tricomi operator
one finds in the works of several authors , among others , agmon @xcite , friedrichs @xcite , gelfand [ unpublished ] , germain and bader @xcite , landau and lifshitz @xcite , leray @xcite , and morawetz @xcite , a large number of examples of solutions to different problems that show the interaction between these two regions .
it seems that the two distributions @xmath31 and @xmath32 are the simplest of such examples .
we now return to equation ( [ eq1c ] ) . in view of the invariance of the tricomi operator under translations parallel to the @xmath3-axis ,
the problem of solving that equation is equivalent to that of solving @xmath33 where @xmath34 is an arbitrary real number and @xmath35 denotes the dirac measure concentrated at the point @xmath36 consider the case @xmath37 , that is , the point @xmath38 is situated in the hyperbolic region .
we introduce the _ characteristic coordinates _ @xmath39 set @xmath40 and to let @xmath41 be such that @xmath42 note that @xmath43 represents the point @xmath38 in characteristic coordinates .
as shown figure 2 in figure 2 , two characteristics with equations @xmath44 pass through the point @xmath36 in characteristic coordinates , these are the two half - lines @xmath45 that originate from the point @xmath46 the characteristics ( [ ch1 ] ) meet the @xmath3-axis , respectively , at the points @xmath30 and @xmath47 two new characteristics originate from these two points , namely , @xmath48 which we call _ reflected characteristics .
_ the corresponding equations in characteristic coordinates are @xmath49 the two characteristics through the point @xmath38 and the reflected characteristics divide the plane into four disjoint regions denoted by @xmath50 @xmath51 @xmath52 and @xmath53 and illustrated in figure 2 . in this paper , we show existence of four fundamental solutions each supported by the closure of the corresponding region . of these solutions ,
only two have physical meaning : the one defined in the region @xmath54 and the other defined in the region @xmath55 this situation is similar to what happens in the case of the wave operator in two dimensions where the two relevant fundamental solutions are the ones supported by the forward and backward light - cones . as we will see , it is the hypergeometric function @xmath0 that plays a crucial role in defining these four fundamental solutions .
the plan of this paper is as follows .
we prove in section [ s2 ] that the function @xmath56 is a solution of @xmath57 where @xmath58 denotes the _ reduced hyperbolic _
tricomi equation ( [ eq3 ] ) . after replacing into ( [ eq6 ] ) , @xmath59 and @xmath60 by their expressions in ( [ ch0 ] )
, we obtain the following function of @xmath3 and @xmath13 : @xmath61^{-1/6}\times\\ & & \phantom{a}\times f\left({1\ov6},{1\ov6};1;{9(x^2-a^2)+4y^3 + 12a(-y)^{3/2}\ov9(x^2-a^2)+4y^3 - 12a(-y)^{3/2}}\right ) .
\nonumber\end{aligned}\ ] ] this is the function that generates the four fundamental solutions relative to the point @xmath36 in section [ s3 ] we define the distribution @xmath62 as the restriction of @xmath63 after multiplication by a suitable constant , to the region @xmath54 and show that @xmath64 is a fundamental solution of the tricomi operator .
the region @xmath54 is entirely contained in the hyperbolic region where it is natural to use characteristic coordinates .
theorem [ th2 ] is then proved , via integration by parts and by using the results of proposition [ pr1 ] . as a consequence of it
we show that , as @xmath38 tends to @xmath65 @xmath62 tends , in the sense of distributions , to @xmath66 the fundamental solution defined by ( [ eq1b ] ) .
the fundamental solutions supported by the closure of the regions @xmath67 and @xmath68 are defined and studied in section [ s4 ] . in both regions , it is necessary to take into account the singularities of @xmath69 along the reflected characteristics .
the results of proposition [ pr2 ] , that describe the asymptotic behavior of the hypergeometric function @xmath0 as @xmath70 are then needed in the proof of theorem [ th3 ] .
we note that as @xmath71 both fundamental solutions @xmath72 and @xmath73 tend to 0 . in section [ s5 ] , we define @xmath74 as the restriction of @xmath63 after multiplication by a suitable constant , to @xmath75 and show that @xmath76 since @xmath77 is complex valued and the tricomi operator has real coefficients , both its complex conjugate and real part are also fundamental solutions relative to the point @xmath36 the imaginary part of @xmath77 is then a solution of the homogeneous equation @xmath78 and we will call it a _ tricomi harmonic function_. contrary to what was proved in corollary [ cor1 ] , it is not true that the fundamental solution @xmath31 defined by ( [ eq1a ] ) is the limit of @xmath79 or its real part , as @xmath80 it is necessary to take a suitable linear combination of @xmath77 and its complex conjugate in order to achieve this result . as a final remark
, we mention that in his paper @xcite , j. leray described a general method , based upon the theory of analytic functions of several complex variables to find fundamental solutions for a class of hyperbolic linear differential operator with analytic coefficients .
in particular , he showed how his method could be used to obtain , in the hyperbolic region , a fundamental solution for the tricomi operator relative to a point @xmath36 he also produced an explicit formula for the fundamental solution in terms of the hypergeometric function @xmath0 .
our method is simpler , more direct , and gives us global fundamental solutions that clearly reflect the change of type of the tricomi operator across the @xmath3-axis .
we would like to thank abbas bahri , fernando cardoso and vladimir retakh for several helpful discussions .
in characteristic coordinates ( [ ch0 ] ) the tricomi operator @xmath82 becomes @xmath83 and we call @xmath58 the _ reduced hyperbolic _ form of @xmath84 we now look for homogeneous solutions of the equation @xmath85 every homogeneous function of @xmath59 and @xmath60 of degree @xmath86 a complex number , can be written as @xmath87 where @xmath88 is a function of a single variable @xmath89 direct substitution into ( [ eq3 ] ) shows that @xmath90 must be a solution of the hypergeometric equation @xmath91\phi'(t ) + { \lambda\over 6}\phi(t)=0.\ ] ] as a solution of this equation , we choose the following hypergeometric function @xmath92 extended , by analytic continuation , to the whole complex plane @xmath93 minus the cut @xmath94 since we are looking for fundamental solutions to the tricomi operator we take for @xmath95 the value @xmath96 as previously indicated in our joint paper @xcite . thus @xmath97 is a solution of @xmath85 let now @xmath98 be an arbitrary point in @xmath99 and consider the change of variables @xmath100 after unenlightening calculations , one can show that the function @xmath101 is also a solution of equation ( [ eq3 ] ) .
this is the special solution to @xmath81 that we are looking for .
consider now the adjoint to equation ( [ eq3 ] ) : @xmath102 one can see that if @xmath103 is a solution of equation ( [ eq5c ] ) , then @xmath104 is a solution of equation ( [ eq3 ] ) .
since @xmath105 is a solution of ( [ eq6 ] ) , it follows that the function @xmath106 is a solution of the adjoint equation ( [ eq5c ] ) .
[ pr1 ] @xmath107 is the unique solution of @xmath108 that satisfies the following conditions : \(i ) @xmath109 along the line @xmath110
\(ii ) @xmath111 along the line @xmath112 ; \(iii ) @xmath113 clearly , conditions @xmath114 imply uniqueness for @xmath115 as @xmath110 the argument of the hypergeometric function in ( [ eq5b ] ) equals zero and so @xmath116 thus , along the line @xmath110 we have that @xmath117 where @xmath118 therefore @xmath119 along @xmath120 in the same manner , one can see that @xmath121 along the line @xmath122 finally , it is clear that @xmath113 @xmath123 the function @xmath107 is the riemann function of the operator @xmath124 relative to the point @xmath125 ( see @xcite ) . if we go back to the notations in the introduction , with @xmath38 and @xmath126 and let @xmath40 then @xmath127 and we obtain from equation ( [ eq5b ] ) the equation ( [ eq6 ] ) that , for further reference , we write as follows : @xmath128 where @xmath129 similarly , after replacing @xmath59 and @xmath60 by their expressions in ( [ ch0 ] ) we obtain the function ( [ eq6 ] ) that we rewrite as @xmath130^{-1/6}\times f\left({1\ov6},{1\ov6};1;\zeta\right ) \nonumber\ ] ] where @xmath131 we wish to analyze this function
. denote by @xmath132 the reflected characteristic @xmath133 at @xmath30 and by @xmath134 the reflected characteristic @xmath135 at @xmath47 [ pr1a ] @xmath69 is a real analytic function defined in @xmath136 the singularities of @xmath69 occur along the reflected characteristics @xmath132 and @xmath137 \1 .
let @xmath138[3(x - a)-2(-y)^{3/2}]=9(x^2-a^2)+4y^3 - 12a(-y)^{3/2}\ ] ] be the denominator of @xmath139 in the expression ( [ ch4a ] ) . for @xmath140 @xmath141 on @xmath142
it is _ negative _ in the region inside these characteristics , and it is _ positive _ in the region outside them . for @xmath143 @xmath144 is a complex number .
it is possible to choose the argument of @xmath144 so that it varies continuously in the region @xmath136 indeed , define @xmath145 the unit circumference , by @xmath146 where @xmath147 since @xmath148 is contractible , @xmath88 lifts to @xmath149 that is , there exists a continuous function @xmath150 defined on @xmath148 so that @xmath151 we may take @xmath150 to be equal zero outside the reflected characteristis and equal @xmath152 inside them .
it follows that @xmath153^{-1/6}\ ] ] is well defined and real analytic in @xmath136 as for @xmath154 it is also real analytic in the same region .
@xmath123 the following remaks are in order : \1 ) for @xmath4 we have @xmath155 and @xmath156 \2 ) for @xmath2 we have @xmath157 it is known ( see @xcite ) that the hypergeometric function @xmath0 is then given by its hypergeometric series and this series is absolutely convergent in the closed disk @xmath158 \3 ) along the characteristics @xmath159 through the point @xmath38 the function @xmath69 is equal to @xmath160 and so it has a singularity of order @xmath161 at @xmath162 \4 ) along both reflected characteristics @xmath134 and @xmath163 , the function @xmath69 has a logarithmic singularity .
this follows from proposition [ pr2 ] , in section [ s4 ] , that describes the asymptotic behavior of the hypergeometric function @xmath0 as @xmath164
let @xmath69 be the function defined by the expressions ( [ ch4 ] ) and ( [ ch4a ] ) and define @xmath165 since @xmath69 is @xmath166 in @xmath54 and bounded on the boundary of @xmath50 it follows that @xmath62 is a locally integrable function and defines a distribution whose support is the closure of @xmath167 [ th2 ] @xmath62 is a fundamental solution for the tricomi operator @xmath82 relative to the point @xmath36 we must show that @xmath168 since @xmath69 is locally integrable in @xmath169 this is equivalent to showing that @xmath170 by introducing the characteristic coordinates @xmath59 and @xmath171 noting that in the new variables the tricomi operator ( [ eq1 ] ) becomes @xmath172,\ ] ] and that the jacobian of the transformation is @xmath173 we can see that the integral in ( [ int1 ] ) is equal to @xmath174 where @xmath175 is @xmath176 in characteristic coordinates .
we denote the last integral by @xmath177 and write it as @xmath178 after recalling that @xmath179 figure 3 next we perform several integrations by parts and take into account the properties of the function @xmath180 with @xmath181 as stated in proposition [ pr1 ] .
first consider the term in ( [ int2 ] ) that contains @xmath182 integrating by parts first with respect to @xmath171 say , and then with respect to @xmath183 one obtains @xmath184 @xmath185 @xmath186 second integrate by parts , relative to @xmath183 the term that contains @xmath187 in ( [ int2 ] ) and obtain @xmath188 @xmath189 finally integrate by parts , relative to @xmath171 the term that contains @xmath190 in ( [ int2 ] ) and get @xmath191 @xmath192 by adding ( [ int3 ] ) , ( [ int4 ] ) , and ( [ int5 ] ) we obtain that @xmath193 and this completes the proof .
@xmath123 as a consequence of theorem [ th2 ] we obtain the following result . [ cor1 ] as @xmath194 the fundamental solution ( [ eq8 ] ) converges in the sense of distributions to @xmath32 the fundamental solution defined by ( [ eq1b ] ) and ( [ eq1b ] ) .
first note that whenever @xmath195 the value @xmath196 is finite and we have @xmath197 thus @xmath198 is well defined . as @xmath194 then @xmath199 and so the limit of @xmath200 defined by ( [ ch3 ] ) and ( [ ch3a ] ) is @xmath201 on the other hand , at the limit the region @xmath54 coincides with @xmath28 the region inside the two characteristics that originate from the origin , where @xmath202 @xmath123 * remarks .
* 1 ) in our joint paper @xcite , the multiplicative constant appearing in the definition of @xmath32 was given by @xmath203 it is a matter of verification that this constant coincides with the one given by formula ( [ eq1b ] ) .
\2 ) theorem [ th2 ] could have been proved in a different way , namely , by using the green s formula ( [ gf ] ) and by integrating along a suitable contour , as we did in our joint paper @xcite .
this method which will be explained in the following section , is more adequate in proving existence of the fundamental solutions supported by the closure of the regions @xmath51 @xmath52 and @xmath204
for reasons of symmetry it is enough to consider only the region @xmath205 this is the region bounded by part of the characteristic @xmath206 part of the characteristic @xmath207 and the reflected characteristic @xmath133 ( see figure 2 ) . as we already pointed out , the function @xmath69 is singular along the reflected characteristic @xmath208 the nature of the singularity depends on the behavior of the hypergeometric function @xmath209 where @xmath210 along the characteristic @xmath211 this is revealed by the following result on the analytic continuation of the hypergeometric series @xmath212 ( see erdly @xcite ) .
the same result also gives us the asymptotic behavior of the analytic extension as @xmath213 [ pr2 ] for @xmath214 with @xmath215 , we have @xmath216\ ] ] where @xmath217 with @xmath218 and @xmath219 with @xmath220 @xmath221 moreover , both series @xmath222 and @xmath223 converge for @xmath224 from this proposition it follows , as we will show in the proof of theorem [ th3 ] , that the singularity of @xmath69 along the reflected characteristics is logarithmic .
we may now define @xmath225 a distribution whose support is the closure of @xmath205 we then have the following result [ th3 ] @xmath226 is a fundamental solution for the tricomi operator @xmath82 relative to @xmath36 1 .
we use the following green s formula for the tricomi operator ( see@xcite ) : @xmath227 where @xmath228 is a bounded domain with smooth boundary @xmath229 if @xmath230 on d , then last formula becomes @xmath231 and the contour integral may be used to evaluate the double integral . now throughout this section the domain @xmath232 lies entirely in the hyperbolic region and so it is more convenient to express the countour integral on the right - hand side of ( [ gf1 ] ) in terms of characteristic coordinates . one can check that , in these coordinates , we have @xmath233 where @xmath234 and @xmath235 similarly , @xmath236 thus the contour integral in ( [ gf1 ] ) , denoted by @xmath237 can be written as @xmath238 figure 4 2 .
we must show that @xmath239 where @xmath240 is the intersection of an open disk that contains the support of @xmath241 and the region , defined in characteristic coordinates , as follows @xmath242 since @xmath243 on the boundary of the disk , it follows from green s formula , that we can replace the double integral in ( [ eq12 ] ) by an integral of the form ( [ ci ] ) , where the contour @xmath244 consists of the oriented line segments @xmath245 @xmath246 and @xmath247 shown in figure 4
. we start by evaluating the simplest of these integrals , namely , the ones along @xmath248 and @xmath249 3 . _ the integral along @xmath250 _
denote by @xmath251 the integral along the line segment , @xmath252 with @xmath60 varying from @xmath253 to @xmath254 taking into account the orientation of @xmath246 we obtain from formula ( [ ci ] ) that @xmath255 where @xmath235 integrating by parts the first integral in ( [ eq13 ] ) , we get @xmath256 @xmath257 @xmath258 from equation ( [ eq5d ] ) it follows that @xmath259 substituting into ( [ eq14 ] ) , we obtain @xmath260 @xmath261 \psi(\ell_0+\epsilon , m)\,dm.\ ] ] now , as @xmath262 the last integral tends to zero because , by proposition [ pr1 ] , @xmath263 along the line @xmath122 on the other hand , the first term in ( [ eq15 ] ) , namely , @xmath264 tends to @xmath265 as @xmath266 therefore , @xmath267 because @xmath268 and @xmath235 4 .
_ the integral along @xmath249 _ denote by @xmath269 the integral along the line @xmath270 with @xmath271 from formula ( [ ci ] ) we now obtain @xmath272 as before , integrate by parts the first integral , use the formula @xmath273 to substitute for @xmath274 in the second integral , and rewrite ( [ eq17 ] ) as @xmath275 @xmath276 \psi(\ell,-\ell_0-\epsilon)\,d\ell.\ ] ] as before , we obtain @xmath277 5 . _
the integral along @xmath278 _ in this case , @xmath279 @xmath59 varies from @xmath280 to @xmath281 and the integral to be considered is @xmath282 our aim is to show that @xmath283 once this is done , then by adding ( [ eq16 ] ) , ( [ eq19 ] ) , and ( [ eq21 ] ) , we obtain ( [ eq12 ] ) and the theorem will be proved . as we did in the previous two cases , we integrate by parts the first integral in ( [ eq20 ] ) and get @xmath284 @xmath285 @xmath286 in order to prove ( [ eq21 ] ) we now must take into account the asymptotic behavior of both @xmath200 and @xmath287 for @xmath60 near @xmath288 and , more specifically , the behavior of the hypergeometric function @xmath0 and its derivative @xmath289 as @xmath290 to see this , we start by expressing ( [ eq22 ] ) in terms of @xmath0 and its derivative . in order to simplify notations we write from now on , @xmath291 and @xmath292 in the case we are considering @xmath293 so
, in view of formula ( [ ch3a ] ) , we set @xmath294 also , recalling formula ( [ ch3 ] ) , we have @xmath295 and so @xmath296 @xmath297 in view of ( [ eq22a ] ) and taking into account that @xmath298 by substituting ( [ eq22b ] ) and ( [ eq22c ] ) into ( [ eq22 ] ) and combining the resulting integrals we obtain @xmath299 @xmath300 @xmath301 as @xmath302 the first term on the right - hand side of ( [ eq23 ] ) clearly tends to zero .
we must prove that the difference of the two integrals in ( [ eq23 ] ) also tends to zero as @xmath303 for this , according to formulas ( [ eq10 ] ) , ( [ eq10a ] ) , and ( [ eq10b ] ) , we have @xmath304 and @xmath305 by using these two expressions we combine the two integrals in ( [ eq23 ] ) and write them as a sum @xmath306 + @xmath307 - @xmath308 where @xmath309 @xmath310 @xmath311 @xmath312 @xmath313 @xmath314 and @xmath315 @xmath316
@xmath317 [ lm1 ] each of the integrals @xmath318 is @xmath319 .
consider first the integral @xmath320 since the series in ( [ eq10a ] ) converge for large values of @xmath321 we may write that @xmath322 where @xmath323 and @xmath324 are bounded functions for small values of @xmath325 by substituting these two expressions into ( [ eq24 ] ) , rewrite it as a sum of three terms : @xmath326 @xmath327 @xmath328 @xmath329 @xmath330 the first term ( inside the brackets ) in ( [ eq27 ] ) tends to @xmath331 as @xmath303 indeed , its limit as @xmath332 is @xmath333 \int_{\ell_0}^\infty{(\ell-\ell_0)^{-5/6}\ov\ell+\ell_0 } \log({2\ell_0(\ell-\ell_0)\ov\ell+\ell_0})\psi(\ell,\ell_0)\,d\ell.\ ] ] the integral on the right - hand side is finite .
on the other hand , @xmath334 are the constant terms of the series @xmath222 and @xmath335 given by ( [ eq10a ] ) .
it is a matter of verification that the quantity inside the brackets is equal zero .
next , consider the second term in ( [ eq27 ] ) .
after the change of variables @xmath336 one can see that the absolute value of that term is bounded above by @xmath337 with @xmath244 a constant independent of @xmath325 with a similar calculation , one can show that the third term in ( [ eq27 ] ) is bounded above by @xmath338 with @xmath244 another constant .
therefore , @xmath339 the expression ( [ eq26 ] ) for @xmath340 similar to that of @xmath341 is even simpler because of the absence of the factor @xmath342 thus , with a similar proof we also conclude that @xmath343 finally , the integral @xmath307 given by ( [ eq25 ] ) is analogous to @xmath344 with @xmath345 and @xmath346 replaced by @xmath103 and @xmath347 the power series defined by ( [ eq10b ] ) .
the proof of the lemma is then complete .
@xmath123 recalling the expression ( [ eq24 ] ) for @xmath348
one sees that lemma [ lm1 ] implies ( [ eq21 ] ) which completes the proof of the theorem .
in the same manner as we did in the previous sections , define the distribution @xmath349 whose support is the closure of the region @xmath55 our aim is to prove the following result .
the proof is , with few modifications , analogous to that of theorem [ th3 ] .
as before , it suffices to show that @xmath350 where the domain of integration @xmath240 is defined as follows .
let @xmath228 be an open disk centered at the origin and with radius sufficiently large so that it contains the points @xmath351 @xmath352 and @xmath38 and the the support of @xmath353 let @xmath354 be the region obtained from @xmath75 by shifting it by @xmath355 along the characteristics and by removing two small half disks centered at the points @xmath351 @xmath352 as shown in the figure 5 .
@xmath240 is then the intersection of @xmath228 and @xmath356 by virtue of the green s formula ( [ gf1 ] ) , the double integral in the expression ( [ eq31 ] ) is to be replaced by a contour integral along the oriented paths @xmath357 @xmath358 @xmath359 @xmath360 @xmath361 and @xmath362 the integral along @xmath360 similar to the integral along @xmath366 ( section [ s4 ] ) , tends to @xmath367 as @xmath303 the same thing happens with the integral along @xmath368 the sum of these two limits is then @xmath369 which is equal to the right - hand side of ( [ eq31 ] ) . recall that @xmath370 to complete the proof one has to show that the limits , as @xmath302 of the integrals along @xmath371 and @xmath372 are both zero . since these two contours lie in the elliptic region of the tricomi operator , it is more convenient to use the _ reduced elliptic _ form of the tricomi operator , namely , @xmath373 which one obtains from equation ( [ eq1 ] ) via the change of variables @xmath374 and @xmath375 and the corresponding green s formula in the variables @xmath3 and @xmath376
. it is then a matter of verification , which we leave to the reader , that along both contours the integrands remain bounded and consequently both integrals along @xmath371 and @xmath372 tend to zero .
the proof is then complete .
we observe that by virtue of proposition [ pr1a ] , the fundamental solution @xmath74 is a complex valued in the upper half plane ( @xmath2 ) plus the region in the lower half plane outside the reflected characteristics and it is real valued in the triangle with vertices @xmath351 @xmath352 and @xmath36 since the tricomi operator has real coefficients , the complex conjugate @xmath377 as well as the real part of @xmath74 are fundamental solutions for @xmath84 contrary to what happened in corollary [ cor1 ] of section [ s4 ] where we showed how the limit of @xmath378 as @xmath194 tends to the fundamental solution @xmath32 given by formulas ( [ eq1b ] ) and ( [ eq1b ] ) , neither of the fundamental solutions just obtained tend to the fundamental solution @xmath31 given by formulas ( [ eq1a ] ) and ( [ eq1a ] ) . for this to happen
we need a particular linear combination of @xmath77 and @xmath379 let @xmath95 and @xmath380 be such that @xmath381 then , the following result holds : j. leray , _ un prolongement de la transformation de laplace qui transforme la solution unitaire dun oprateur hyperbolique en sa solution lmentaire ( problme de cauchy , iv ) _ , bull .
france , * 90 * , fasc .
i , 39156 ( 1962 ) . | in this paper we explicitly calculate fundamental solutions for the tricomi operator , relative to an arbitrary point in the plane , and show that all such fundamental solutions originate from the hypergeometric function @xmath0 that is obtained when we look for homogeneous solutions to the reduced hyperbolic tricomi equation . |
the mg@xmath1 3s-3p fine structure doublet near 280 nm ( see fig . [
fig : mg_levels ] ) is a prominent feature in many astronomical spectra .
it can be observed by ground based telescopes , once the cosmological redshift translates the strong uv lines into the transparency range of the atmosphere .
this makes the transitions good probes of physics at early epochs .
one example for their astrophysical importance is the ongoing quest for a cosmological space - time variation of the fine structure constant @xmath3 .
an investigation of narrow absorption lines from gas clouds which are back - illuminated by distant quasars ( quasi - stellar objects , qso ) , allows to compare the value of @xmath3 in distant regions of the universe with the value measured on earth @xcite .
the detection of a variation of the electromagnetic coupling constant @xmath3 would provide compelling evidence of physics beyond the standard model .
the mg@xmath1 doublet was first used in 1976 @xcite to constrain variations of @xmath3 via the alkali doublet method @xcite .
an analysis of fine structure doublets allows to distinguish between a variation in @xmath3 and a frequency change due to the cosmological redshift , since the ratio between the fine structure and gross structure of atomic energy levels scales as @xmath4 in leading order .
dzuba et al .
introduced the many multiplet ( mm ) method in 1999 @xcite , a refined approach which utilizes the fact that the transition frequencies itself contain relativistic corrections @xmath5 .
the relativistic corrections are most pronounced for the ground states , so the comparison of transitions in different atomic species from the same object can significantly improve the sensitivity .
resonance doublet near 280 nm is isoelectronic to the 3s-3p doublet in na historically labeled as d@xmath6 and d@xmath7 fraunhofer lines . from the d@xmath6 and d@xmath7 transition frequencies measured in @xmath0mg@xmath1 and @xmath2mg@xmath1 we deduce isotope shifts and fine structure splittings ( fs ) .
the measurement accuracy allowed to resolve the isotope shift of the fine structure splitting for the first time . ]
webb and coworkers applied the mm method to fe@xmath1 transitions measured against the mg@xmath1 doublet and found indications that @xmath3 was smaller than the present day value for redshifts @xmath8 ( about @xmath9 years ago , @xmath10 ) @xcite .
the same group extended their studies to larger samples and further transition lines @xcite , which finally resulted in a statistically significant signature of a smaller fine structure constant @xmath11 .
hereby , 143 qso absorption samples where evaluated ( 77 of which were low-@xmath12 mg / fe systems @xcite ) , all observed by the keck observatory using the hires spectrometer @xcite .
attempts to reproduce the results with 23 mg - based samples observed by a different telescope and spectrometer ( vlt / uves @xcite ) suggest a drift consistent with zero @xmath13 @xcite . this work and further investigations consistent with zero drift
are summarized and disputed in @xcite .
recent single ion differential @xmath3 measurement ( sidam )
studies on fe@xmath1 lines also gave a result compatible with zero drift @xcite .
next generation telescopes @xcite , possibly in combination with absolute frequency calibration @xcite may help to clarify the discrepancies on the reported uncertainty levels .
comparison with qso absorption spectra requires laboratory reference data with an accuracy of better than 10@xmath14 @xcite , which corresponds to about 40 mhz at 280 nm .
absolute frequency calibrated measurements @xcite in the favorable environment of an ion trap , as reported in this article , provides reference data @xcite that may be regarded as exact for comparison with astrophysical spectra . besides high accuracy it is particularly important to provide isotopically resolved data .
magnesium has three stable isotopes with mass numbers 24 , 25 and 26 ( natural abundance 79:10:11 ) .
different isotope abundances in the investigated astronomical objects were identified as an important source of systematic errors in the mm analysis , moreover , an investigation of the evolution of isotope abundances itself is of interest @xcite . for these reasons an isotope shift measurement on the d@xmath6 transition
was requested in @xcite .
also comparison with isotope shift theory demands refined reference data since the accuracy of theoretical predictions exceeded the measurement uncertainty @xcite . so far the only isotopically resolved measurement was provided by drullinger et al .
@xcite , where the @xmath15mg@xmath1 isotope shifts of the d@xmath16 line were determined with an uncertainty of 100 mhz .
the most accurate literature values for the d@xmath17 and d@xmath16 transition frequencies originate from fourier transform ( ft ) spectrometry using a hollow - cathode discharge lamp here an uncertainty of 60 mhz was reported for unresolved lines of natural isotopic abundance @xcite .
mm analysis used laboratory reference data for the individual isotopic ( and hyperfine structure ) components @xcite that are derived from the unresolved ft measurements @xcite in combination with the d@xmath16 isotope shift results @xcite .
our measurement of the d@xmath17 and d@xmath16 component of the @xmath0mg@xmath1 and @xmath2mg@xmath1 isotopes , improves the transition frequencies by more than two orders of magnitude . from these four absolute frequency measurements we derive @xmath18mg@xmath1 isotope shifts .
the isotope shift of the d@xmath17 transition has not been measured before and the d@xmath16 value is improved by three orders of magnitude .
the accuracy of our measurement allows to resolve the @xmath18mg@xmath1 isotope shift of the fine structure splitting . from the @xmath0mg@xmath1 and @xmath2mg@xmath1 measurements we additionally deduce @xmath19mg@xmath1 transition frequencies , including the hyperfine structure , with an accuracy that is better than requested in @xcite .
finally , we note that strong constraints on _ present - day _ variations of @xmath3 are set terrestrially by a comparison of optical clocks based on different atomic transitions .
though optical clocks may only be compared on time scales of years , the tremendous accuracy allows to constrain variations of @xmath3 about an order of magnitude more accurately than mm studies , if a linear drift over 10@xmath20 years look back time is assumed .
recent measurements of an al@xmath1 clock against a hg@xmath1 clock at nist suggested @xmath21/year @xcite .
mg@xmath1 d@xmath6 resonance .
fluorescence photons were collected for 3s per data point , the scan was randomized in order . for our measurements we restricted the tuning range to 180 mhz around the line center .
inset : for comparison we show an asymmetric @xmath0mg@xmath1 d@xmath7 line profile scanned from the low frequency side , overlayed is a voigt fit to the rising edge , which is centered around the measured transition frequency . here
the ion was essentially uncooled , only a weak cooling beam was superimposed to prevent ion loss during the scan time . ] lines observed in qso absorption spectra are usually dipole - allowed transitions from the ground state @xcite .
spectroscopy on such transitions in trapped ions requires different spectroscopic approaches than probing narrow clock transitions .
spectroscopy in the favorable sideband resolved regime @xcite , that allows to avoid first order doppler and recoil shifts , is hard to establish , since typical transition bandwidths @xmath22 are broad compared to practical secular trap frequencies @xmath23 . in the weak binding limit ( @xmath24 )
the absorption spectrum is strongly asymmetric if no further measures are taken and the transition is probed with a single laser beam . in this regime
the motional sidebands overlap and for constant temperature a voigt profile would be observed , just as for an unbound atom @xcite .
however , scattering photons changes the temperature of the ion ; on the low - frequency ( `` red '' ) side of the resonance the ion is cooled to a detuning dependent temperature , on the high - frequency ( `` blue '' ) side it is heated and expelled from the trap center . therefore the voigt spectrum is heavily distorted : on the red side due to a detuning dependent equilibrium temperature and thus varying gaussian contribution to the linewidth .
near resonance the temperature diverges and the ion is lost , so the blue side is not observed @xcite ( see also fig . [
fig : widescan2 ] inset ) .
therefore , in view of precision spectroscopy , the crux is to suppress adverse temperature variations induced by the spectroscopy laser .
since the statistical uncertainty determining the line center is given by the linewidth divided by the signal - to - noise ratio of the data , a suitable approach should ensure low temperature ( i.e. narrow linewidth ) and a high scattering rate .
previously used approaches include a two laser method @xcite , a double resonance scheme @xcite , buffer gas cooling @xcite and chopped detection @xcite .
our novel spectroscopy technique allows to record essentially unperturbed voigt profiles for the first time .
the method is discussed in detail in @xcite , here we will only sketch the basic idea .
we prepare a crystallized chain of ions which is continuously doppler cooled at one side only ( see fig . [ chain ] ) .
a weaker spectroscopy laser probes sympathetically cooled ions at the other end of the chain , where we collect a basically background free fluorescence signal from single ions using an imaging photon detection system .
this avoids both the background and ac stark shift from the cooling laser .
counterintuitive to a thermal conductivity picture , the symmetric eigenmode structure efficiently couples cooling and spectroscopy ions at opposite ends of the chain , i.e. the sympathetically cooled ions are cooled just as efficiently as directly cooled ions .
a detailed analysis of the considered transitions shows : residual temperature variations lead to systematic shifts of the line center always below @xmath25 of the linewidth , as long as the cooling laser damping exceeds the heating induced by the spectroscopy beam for all detunings .
therefore , the method enables to observe virtually unperturbed resonances from single , cold ions with high signal - to - noise ratio ( see fig . [
fig : widescan2 ] ) .
the experimental setup is described in @xcite , here we summarize basic features and illuminate further aspects in greater detail .
the mg@xmath1 ions were stored in a linear paul trap operated at 15.8 mhz .
the radial secular frequencies were about 1 mhz , while the axial secular frequency was chosen to be about 60 khz .
ions were loaded into the trap by two - step photo ionization via the 3s@xmath26 - 3p@xmath6 transition in neutral mg with a frequency doubled dye laser at 285 nm ( coherent 699 ) .
references @xcite showed that isotope selective photo ionization is possible when atom and laser beams are carefully arranged .
our trap was loaded from neutral mg vapor of natural abundance ( doppler broadening : about 3 ghz , @xmath18 mg 3s@xmath26 - 3p@xmath6 isotope shift : 1.4 ghz @xcite ) .
yet we observed purified chains of up to 30 @xmath2mg@xmath1 ions , when we applied additional radio frequencies resonant to the @xmath0mg@xmath1 and @xmath19mg@xmath1 radial secular frequencies to an electrode parallel to the trap axis during the loading period .
cooling and spectroscopy light was either provided by another frequency doubled dye laser ( coherent 699 ) or a frequency quadrupled yb fiber laser system @xcite at 1118 nm ( koheras boostik ) , which could only interact with the d@xmath7 transition ( 90 ghz tuning range ) . a schematic setup using both lasers is shown in fig .
[ fig : setup ] . after a failure of the fiber laser
the cooling and spectroscopy beams were produced by the frequency doubled dye laser . in this case
the uv output was split and passed through two double pass acousto - optic modulators ( aom ) setups @xcite , which allowed to independently control the frequency and amplitude of both beams .
the intensities were set to @xmath27 and @xmath28 in units of the saturation intensity ( @xmath29 2.47 mw / mm@xmath30 ) .
the cooling beam was detuned 64 mhz below resonance , the practical tuning range of the spectroscopy laser was 180 mhz , limited by the efficiency of the scanning aom in double pass configuration .
the cooling beam was tightly focused to a beam radius of @xmath31 m and enclosed an angle of about 15@xmath32 with the trap axis predefined by the optical access of our vacuum vessel . given that geometry we found optimal experimental conditions when we cooled 2 - 3 outer ions of a chain containing 8 - 10 ions .
this allowed us to perform low background spectroscopy on 2 - 3 ions at the other end of the chain . working with shorter chains decreased the number of low background spectroscopy ions , longer chains increased the probability that a mgh@xmath1 ion was generated by a chemical reaction with the background gas .
dark mgh@xmath1 ions embedded in the chain changed their site on time scales of few seconds . once they enter the cooling or spectroscopy region it turns dark which distorts the line profile .
we therefore reloaded our trap once an impurity ion was observed .
the spectroscopy beam collinear with the ion chain was stabilized in intensity , spatially filtered with a small pinhole and loosely focused to about @xmath33 m in the trapping region .
additional spatial filtering was necessary since tuning of the double pass aom led to frequency dependent distortions of the beam profile , observable with a beam profiling camera .
we stabilized the _ total _ intensity , but the ions sampled frequency dependent _ local _ intensity variations , which resulted in an apparent alignment sensitivity of the line center .
filtering with a 15 @xmath34 m size pinhole removed the beam profile distortions and the alignment sensitivity of the line center .
the rf reference for all involved frequency combs , phase locked loops and the spectroscopy aom was provided by a hydrogen maser .
a comparison between the free - running maser and the global positioning system showed a deviation smaller than @xmath35 over the whole measurement period , which is sufficient for our purpose .
the phase coherent link between the rf reference and the optical regime was achieved with a 100 mhz repetition rate erbium doped fiber laser frequency comb .
the frequency comb reaches well into the fourth sub - harmonic of the mg@xmath1 doublet at about 1120 nm . to bridge the gap to the green dye laser output we phase locked a diode laser at 1120 nm to a mode of the frequency comb and recorded a beat note between its frequency doubled output and the dye laser ( see fig .
[ fig : setup ] ) .
a synthesizer mixed the beat frequency down to a constant offset lock frequency , which allowed us to arbitrarily set the frequency of the phase locked dye laser with respect to the ir comb modes .
this technique was also used to tune the spectroscopy laser , when we cooled the ion chain with the independent fiber laser system .
the mode number of the ir comb , used for the measurement of the @xmath0mg@xmath1 d@xmath7 line was determined with a second frequency comb with repetition rate of 250 mhz ( menlo systems ) @xcite .
we deduced the mode number for the remaining transitions relative to the known @xmath0mg@xmath1 d@xmath7 component using two independent wavemeters with an relative accuracy considerably better than the repetition rate ( burleigh wa 1500 , high finesse ws7 ) .
every single measurement consisted of 30 - 40 frequency settings within a 180 mhz ( @xmath36 4 @xmath37 ) wide range around the line center , fluorescence photon counts were recorded for 3 s each .
the chronological order of probe laser frequencies was randomized for each line scan to be insensitive to a great variety of possible drifts .
photons were collected with a f/2 imaging system attached to a single photon camera ( quantar mepsicron ii ) .
the spatially resolved count rate was digitized in 512 @xmath38 512 pixel images . to read out the fluorescence of a single ion we set a circular region of interest around the ion using software ( see fig .
[ chain ] ) . [ cols="<,^,^,^",options="header " , ] [ tableshift ] the accuracy of theoretical isotope shifts used to exceed experimental data @xcite , our measurement reverses this situation .
theoretical isotope shifts from recent investigations @xcite are shown in table [ tableshift ] .
different results arise mainly from the determination of the sms , since differences in the nms and differences in the field shift due to uncommon field shift constants @xmath39 are marginal , compared to the uncertainty introduced by experimental rms nuclear charge radii @xcite .
only the authors of @xcite give error bars for their evaluation of the sms constant _
k_@xmath40 , the uncertainty levels introduced by _
k_@xmath40 and the nuclear charge radii are both on the order of 20 mhz .
therefore the considerably more precise value of our measurement can neither set an accurate calibration point for sms investigations nor give rigorous insight on isotopic differences in the nuclear charge distribution .
yet the results of @xcite and our measurement are in good agreement . combining our results with the values given for _
k_@xmath40 and @xmath39 from @xcite allows to deduce the difference in the rms nuclear charge radii .
here we find @xmath41= 0.06 ( 15 ) fm@xmath30 from the d@xmath7 line and @xmath41= 0.07 ( 30 ) fm@xmath30 from the d@xmath6 result , consistent with the tabulated value of 0.14 ( 16 ) fm@xmath30 @xcite .
the accuracy of our measurement allows to resolve the @xmath18mg@xmath1 isotope shift of the forbidden 3p@xmath43 - 3p@xmath44 transition .
since the fine structure splitting is of purely relativistic origin our measurement may be an interesting test for relativistic effects in isotope shift investigations .
one result of @xcite was , that the final relativistic mass shifts for levels of a multiplet appear to be closer to each other than the mass shifts obtained with non relativistic operators , giving rise to a difference in the determined mass shifts of 2.2 mhz and -14.4 mhz for the relativistic and non relativistic treatment respectively ( see table v in @xcite ) .
our measured value of 2.66 ( 13 ) mhz for the total isotope shift supports this result .
we have accurately measured the @xmath0 mg and @xmath2 mg components of the 3s-3p fine structure doublet in mg@xmath1 using a new spectroscopy method for trapped ions in the weak binding limit @xcite .
absolute frequency calibration ( relative to the definition of the si second ) allows us to directly deduce the corresponding isotope shifts and fine structure splittings .
the measurement accuracy exceeds previous results by more than two orders of magnitude and allows to resolve the isotope shift of the relativistic fine structure splitting for the first time .
we deduce transition frequencies for the @xmath19 mg hyperfine structure and the center - of - mass of an isotope mixture of natural abundance .
a further measurement of the @xmath19 mg component would improve the uncertainty of composite transition frequencies , isotope shifts and hyperfine constants for the 3p@xmath44 and 3p@xmath43 level @xcite . in turn
it would allow for a more stringent consistency check of theoretical hyperfine structure constants , isotope shift theories and nuclear charge distributions .
though our measurement may be regarded as exact reference for present and near future astrophysical spectra and theoretical isotope shifts , future applications may require even more accurate calibration .
narrowing down the linewidth of the spectroscopy laser e.g. by stabilization to an ultra - stable reference cavity would essentially remove the limiting systematic effect of our measurement .
the statistical uncertainty of the @xmath2d@xmath7 line was only 14 khz , indicating that our spectroscopy scheme may allow for further improvement of at least one order of magnitude .
t. rosenband , d. b. hume , p. o. schmidt , c. w. chou , a. brusch , l. lorini , w. h. oskay , r. e. drullinger , t. m. fortier , j. e. stalnaker , s. a. diddams , w. c. swann , n. r. newbury , w. m. itano , d. j. wineland , and j. c. bergquist , science * 319 * , 1808 ( 2008 ) . | we apply a recently demonstrated method for precision spectroscopy on strong transitions in trapped ions to measure both fine structure components of the 3s-3p transition in @xmath0mg@xmath1 and @xmath2mg@xmath1 .
we deduce absolute frequency reference data for transition frequencies , isotope shifts and fine structure splittings that are in particular useful for comparison with quasar absorption spectra , which test possible space - time variations of the fine structure constant .
the measurement accuracy improves previous literature values , when existing , by more than two orders of magnitude . |
the electrical noise spectrum in quantum systems is accessible experimentally through very sensitive techniques such as spectrum analyzer @xcite , use of a superconductor - insulator - superconductor tunnel junction as an on - chip spectrum analyzers @xcite and measurement of photon emission spectrum @xcite .
it exists also a proposal to measure the heat statistics in quantum devices @xcite .
given the fast progress of detection techniques , it is not forbidden to imagine that in the next few years , the measurement of the noise correlator between the electrical current and the heat current ( mixed noise ) would be possible . in parallel ,
calculations of finite frequency electrical / heat mixed noise are needed for quantum systems . there
exist very few works on the zero - frequency mixed noise @xcite and not even one concerning the finite - frequency mixed noise .
theoretically , the studies are limited to the electrical noise spectrum ( see @xcite and references therein ) , to the energy noise spectrum @xcite , to the statistics of the energy current in the presence of time - dependent excitation @xcite , and to the heat noise spectrum @xcite .
this is regrettable since it has been shown recently that the zero - frequency mixed noise contains information on the thermoelectric response of the system @xcite : it gives the figure of merit in the linear response regime and it is related to the thermoelectric efficiency in the weak transmission regime ( schottky regime ) . at finite - frequency , the mixed noise should bring information on the dynamics of the thermoelectric conversion , in particular on the thermoelectric response of time - modulated systems , which is the study of an increasing number of works @xcite . in this paper
, we fill this lacuna by calculating the mixed noise spectrum of a quantum dot ( qd ) using the keldysh out - of - equilibrium green function technique .
we focus on the non - symmetrized noise spectrum since this is the quantity which is relevant for quantum systems , due to the fact that the current operators do not commute with each other @xcite .
the paper is organized as follows : we present the model and give the definition of electrical and heat currents in sec .
the results for the noise spectra are presented in sec .
iii , and discussed in sec .
we conclude in sec .
to model the qd connected to left ( l ) and right ( r ) reservoirs , we use the hamiltonian @xmath0 , where @xmath1 describes the energy of electrons in the reservoir @xmath2 , with @xmath3 ( @xmath4 ) , the creation ( annihilation ) operator , @xmath5 describes the qd with a single energy level @xmath6 , with @xmath7 ( @xmath8 ) the creation ( annihilation ) operator , and @xmath9 describes the transfer of electrons from the reservoirs to the qd and vice versa .
the left and right reservoirs are assumed to be at equilibrium with temperature @xmath10 and chemical potential @xmath11 ( see fig .
[ figure1 ] ) . and @xmath12 ) .
the black arrows indicate the convention chosen for the currents direction.,width=264 ] the charge current , @xmath13 , and heat current , @xmath14 , flowing from the reservoir @xmath2 to the qd , are given by the time derivatives @xcite of the operators _ number _ of electrons in the reservoir @xmath2 , @xmath15 , and _ energy _ of electrons in the reservoir @xmath2 , @xmath16 : @xmath17 , and @xmath18 , with @xmath19 .
the time derivatives of these two quantities are equal to @xmath20 $ ] , and @xmath21 $ ] , which lead after calculation to @xmath22 injecting eqs .
( [ n ] ) and ( [ h ] ) in the definitions of @xmath13 and @xmath14 , we obtain : @xmath23 which give in a compact form @xmath24
the non - symmetrized noise spectrum is defined as @xmath25 where @xmath26 is the current - current time - correlator , and @xmath27 , with @xmath28 , the electrical ( @xmath29 ) or heat ( @xmath30 ) current operator from the reservoir @xmath2 to the central region through the barrier @xmath2 . the finite - frequency non - symmetrized noise @xmath31 quantifies the correlation between the currents @xmath32 and @xmath33 at finite frequency @xmath34 .
such a general definition embeds three types of noises : ( i ) the charge noise , @xmath35 , corresponding to the correlator between the electrical current and itself ; ( ii ) the mixed noises , @xmath36 and @xmath36 , corresponding to the correlators between the electrical current and the heat current , and ( iii ) the heat noise , @xmath37 , corresponding to the correlator between the heat current and itself .
we first compute the time - correlator @xmath38 using the keldysh out - of - equilibrium formalism @xcite , and next calculate its fourier transform in order to get @xmath31 . to achieve this task , we insert the current operator , given by eq .
( [ c1 ] ) , in the definition of the noise , given by eq .
( [ def_s ] ) , and perform the calculation of the average of the product of four creation / annihilation operators , making the following assumptions : non - interacting electrons , wide - band approximation and symmetrical coupling strength between the reservoirs and the qd ( i.e. , symmetrical left and right barriers ) .
the details of the calculation are given in the appendix [ appendixa ] .
the final expression of finite - frequency non - symmetrized noise we obtain is @xmath39~,\end{aligned}\ ] ] with @xmath40f_\alpha^e(\varepsilon)+[\delta_{\alpha\beta}-t^*(\varepsilon)]f_\beta^e(\varepsilon)\big]~,\end{aligned}\ ] ] @xmath41\nonumber\\ & & \times \left[f_\beta^h(\varepsilon-\hbar\omega)-t^*(\varepsilon-\hbar\omega)f_m^h(\varepsilon-\hbar\omega)\right]~,\end{aligned}\ ] ] and @xmath42f_\alpha^h(\varepsilon-\hbar\omega)\nonumber\\ & & + [ \delta_{\alpha\beta}-t(\varepsilon-\hbar\omega)]f_\beta^h(\varepsilon-\hbar\omega)\big]~,\end{aligned}\ ] ] where @xmath43^{-1}$ ] is the fermi - dirac distribution function for electrons , @xmath44 is the distribution function for holes , @xmath45/2 $ ] is the average left and right distribution , @xmath46 is the transmission amplitude , and @xmath47 is the transmission coefficient . the transmission amplitude is related to the retarded green function of the qd , @xmath48 , through the relation @xmath49 , where @xmath50 is the coupling strength between the qd and the reservoirs @xcite .
equation ( [ exp_noise ] ) gives the electrical noise when @xmath51 , it gives the mixed noise when either @xmath29 and @xmath52 , or vice versa , and it gives the heat noise when @xmath53 .
@xmath31 is a real quantity when @xmath54 and @xmath55 ( auto - correlator ) , but can be complex otherwise ( cross - correlator ) .
we have checked that the electrical noise @xmath35 extracted from eq .
( [ exp_noise ] ) coincides with the results of the literature @xcite , and that the heat noise @xmath37 extracted from eq .
( [ exp_noise ] ) coincides with the existing results of the literature in the limit of energy independent transmission amplitude @xcite .
the expressions for the mixed noises @xmath36 and @xmath56 are novels .
this is the central result of this paper .
it is valid at any frequency @xmath34 , coupling strength @xmath50 between the qd and the reservoirs , qd energy level @xmath6 , and for any temperature and voltage gradients between the left and right reservoirs . in the following ,
we choose first to restrict our study to the case where temperatures for the left and right reservoirs are equal , @xmath57 , and for @xmath58 , and we discuss the mixed noise spectrum in three situations : ( i ) at equilibrium , ( ii ) for energy independent transmission amplitude , and ( iii ) for an anderson - type transmission amplitude . in the latter case , we also discuss the spectral coherence in the presence of a temperature gradient between the two reservoirs .
at equilibrium , i.e. , @xmath59 , where @xmath60 is the fermi energy for electrons in the reservoirs , and for equal left and right reservoir temperatures , i.e. , @xmath57 , we have from eqs .
( [ exp_noise]-[c ] ) @xmath61~,\nonumber\\\end{aligned}\ ] ] with @xmath62~,\\ \mathcal{\widetilde b}_{\alpha\beta}(\varepsilon,\omega ) & = & t(\varepsilon)t(\varepsilon-\hbar\omega)\nonumber\\ & & \times\left[1-t^*(\varepsilon)\right]\left[1-t^*(\varepsilon-\hbar\omega)\right]~,\\ \mathcal{\widetilde c}_{\alpha\beta}(\varepsilon,\omega ) & = & \mathcal{t}(\varepsilon ) \left[2\delta_{\alpha\beta}-\mathcal{t}(\varepsilon-\hbar\omega)\right]~,\end{aligned}\ ] ] since for isothermal reservoirs at equilibrium , we have @xmath63 . from eq .
( [ s_eq ] ) , it can be shown using @xmath64 that the noise spectrum obeys the relation @xmath65~,\end{aligned}\ ] ] where @xmath66^{-1}$ ] is the bose - einstein distribution function . removing the reservoirs index and using the definitions @xcite of the electrical ac - conductance , @xmath67/2\hbar\omega$ ] , the thermal ac - conductance @xmath68/2\hbar\omega t$ ] , and the thermoelectric ac - conductance , @xmath69/2\hbar\omega t$ ] which is the product of the ac - thermopower ( i.e. , seebeck coefficient ) by the electrical ac - conductance
, we establish that the noises are related at equilibrium to the ac - conductances through the following fluctuation - dissipation relations @xmath70 through these relations , we can state that in a similar way that the finite - frequency electrical noise contains information on the dynamics of the charge transfer , the finite - frequency heat noise contains information on the dynamic of the heat transfer ( since @xmath71 and @xmath72 are the response to an excitation modulated in time ) .
moreover , eq . ( [ x ] ) confirms the key role played by the mixed noise @xmath73 to quantify the thermoelectric conversion .
note that in the limit of zero - frequency , eqs .
( [ g]-[k ] ) reduce to the relations given in ref .
, since we have in that limit @xmath74 . for an energy - independent transmission amplitude , @xmath75 ,
the real parts of the electrical , mixed and heat noise spectra are given by fig .
[ figure2 ] in the low temperature limit .
we do not plot their imaginary parts whose magnitudes are smaller with a factor hundred comparing to the ones of the real parts .
let us now discuss the features appearing on fig .
[ figure2 ] .
first , we notice that similarly to the electrical noise , which cancels when the frequency is larger than the voltage , @xmath76 , the mixed and heat noises cancel as well .
the reason is the following : knowing that the noise is called emission noise at positive frequency and absorption noise at negative frequency @xcite , we understand that the system can not emit an energy larger than the energy provided to it , here the voltage since temperature is taken small .
second , we observe that the electrical noise varies linearly or by plateaus with both voltage and frequency , due to the fact that when transmission is energy independent , the system works in the linear regime .
third , the mixed noise can change its sign whereas the electrical and heat noises keep a single sign .
fourth , the electrical and mixed correlators between distinct reservoirs , @xmath77 , are equal in absolute values to the correlators in the same reservoir , @xmath78 , and nearly equal for the heat correlator @xcite . in the limit of weak or perfect transmission , i.e. , @xmath85 or @xmath86 respectively , the integration over energy in eq .
( [ exp_noise ] ) can be performed analytically ( see appendix [ appendixb ] for the details of the calculation ) .
the expressions of the noises , which are all real in these limits , are given in table [ table1 ] .
these expressions constitute a generalization of the fluctuation - dissipation theorem to an out - of - equilibrium situation since the bose - einstein distribution function is estimated at frequency shifted by @xmath87 .
concerning the electrical noise , its expression at @xmath85 is in full agreement with the result of perturbative calculations @xcite . concerning the mixed and heat noises expressions
, there is no previous work to compare in the literature .
note that at zero - voltage , the mixed noise cancels in both limits ( @xmath85 and @xmath86 ) but not in the intermediate regime : the mixed noise is then given by eq .
( [ x ] ) which is a priori non - zero .
it is also worth to notice that @xmath37 contains a contribution which is proportional to @xmath35 with a proportionality factor equal to @xmath88 , where @xmath89 is the lorenz number . since in certain limits ,
the heat noise is related to the thermal conductance and the electrical noise to the electrical conductance , as through eqs .
( [ k ] ) and ( [ g ] ) at equilibrium for example , it is not surprising to find a relation which involves the lorenz number between the heat noise and the electrical noise thanks to the wiedemann - franz law , or between the thermal conductance and the electrical noise as obtained in ref . .
table [ table1 ] gives also the sum over reservoirs of the electrical , mixed and heat noises , @xmath90 .
contrary to the total electrical and mixed noises , which are equal to zero in the limits we consider ( no charging effect on the qd ) , the total heat noise takes a finite value which indeed corresponds to the heat power fluctuations . at zero - frequency , the power fluctuations is conserved , i.e. , the heat power fluctuations is equal to the electrical power fluctuations @xcite .
it is also true at finite - frequency provided that @xmath86 , since in that limit we have from table [ table1 ] : @xmath91 . at zero - temperature ,
we get for @xmath86 : @xmath92 , and for @xmath85 , @xmath93 at zero - voltage , and @xmath94 at zero - frequency in good agreement with ref . .
.expressions of the electrical , mixed and heat noises in the energy independent weak / perfect transmission limits @xcite .
we have @xmath95 when @xmath96 , and vice versa .
the total electrical , mixed and heat noises summed over both reservoirs are also given .
the details of the calculations are performed in appendix b. [ cols="<,^,^,^",options="header " , ] in the following sections , we will meet the integral @xmath97 with @xmath98 = 0 , 1 or 2 . here
, we calculate this integral considering the isothermal case , @xmath99 , thus @xmath100 where we have introduced the bose - einstein distribution function : @xmath66^{-1}$ ] . using the identity : @xmath101=\tanh(a)-\tanh(b)$ ]
, we end up with @xmath102~.\nonumber\\\end{aligned}\ ] ] to go further , we perform a taylor expansion up to the third order with @xmath103 , @xmath104 or @xmath105 .
it leads to @xmath106~,\nonumber\\\end{aligned}\ ] ] with @xmath107 $ ] , @xmath108 $ ] , and @xmath109 $ ] .
the calculation of these integrals gives @xmath110 @xmath111 and @xmath112 we calculate only @xmath114 since when @xmath75 and @xmath115 are independent of energy , we have the relations @xmath116 at weak @xmath115 , we have @xmath117d\varepsilon\nonumber\\ & = & \frac{e^{2}\mathcal{t}}{h}\big [ i^{(0)}_{lr}+ i^{(0)}_{rl}\big]~,\end{aligned}\ ] ] which gives @xmath118~.\end{aligned}\ ] ] it reduces at equilibrium ( zero - voltage ) to @xmath119 and at zero - temperature to @xmath120~,\end{aligned}\ ] ] since we have @xmath121 when @xmath122 .
we start from @xmath123f^e_\gamma(\varepsilon)f^h_\delta(\varepsilon-\hbar\omega)~.\end{aligned}\ ] ] we calculate only @xmath124 since at weak @xmath115 , we have : @xmath125 we have @xmath126-\hbar\omega f^e_r(\varepsilon)f^h_l(\varepsilon-\hbar\omega)\big]\nonumber\\ & = & \frac{e\mathcal{t}}{h}\big[i^{(1)}_{lr}-\mu_l i^{(0)}_{lr}+i^{(1)}_{rl}-(\mu_l+\hbar\omega)i^{(0)}_{rl}\big]~,\nonumber\\\end{aligned}\ ] ] which gives @xmath127~.\end{aligned}\ ] ] it reduces to @xmath128 at equilibrium ( zero - voltage ) and to @xmath129~,\end{aligned}\ ] ] at zero - temperature .
we start from @xmath130f^e_\gamma(\varepsilon)f^h_\delta(\varepsilon-\hbar\omega)~,\end{aligned}\ ] ] which gives for @xmath131 @xmath132 - 2\hbar\omega(\varepsilon-\mu_l)f^e_r(\varepsilon)f^h_l(\varepsilon-\hbar\omega)\nonumber\\ & & + \hbar^2\omega^2[f^e_l(\varepsilon)f^h_l(\varepsilon-\hbar\omega)+f^e_r(\varepsilon)f^h_l(\varepsilon-\hbar\omega)]\big]\nonumber\\ & & = \frac{\mathcal{t}}{h } \bigg[i^{(2)}_{lr}+i^{(2)}_{rl}-2\mu_l(i^{(1)}_{lr}+i^{(1)}_{rl})+\mu_l^2(i^{(0)}_{lr}\nonumber\\ & & + i^{(0)}_{rl})-2\hbar\omega(i^{(1)}_{rl}-\mu_li^{(0)}_{rl } ) + \hbar^2\omega^2(i^{(0)}_{ll}+i^{(0)}_{rl})\bigg]~.\nonumber\\\end{aligned}\ ] ] we report the expressions of the integrals given by eqs .
( [ i0]-[i2 ] ) and factorize the various contributions .
it gives @xmath133\nonumber\\ & & + \frac{(\hbar\omega - ev)^3}{3}n(\hbar\omega - ev)\nonumber\\ & & + \frac{(\hbar\omega+ev)^3}{3}n(\hbar\omega+ev)\bigg]~,\end{aligned}\ ] ] which reduces at equilibrium ( zero - voltage ) to @xmath134n(\hbar\omega)~,\nonumber\\\end{aligned}\ ] ] and at zero - temperature to @xmath135~,\end{aligned}\ ] ] since we have @xmath121 when @xmath122 .
note that @xmath136 is obtained from the expression of @xmath131 by inverting the voltage @xmath137 , as a consequence we have @xmath138 .
we now calculate @xmath139\nonumber\\ & & + \hbar\omega(\varepsilon-\mu_l)f^e_l(\varepsilon)f^h_r(\varepsilon-\hbar\omega)\nonumber\\ & & + \hbar\omega(\varepsilon-\mu_r)f^e_r(\varepsilon)f^h_l(\varepsilon-\hbar\omega)]\big]\nonumber\\ & = & \frac{\mathcal{t}}{h}\big[-i^{(2)}_{lr}-i^{(2)}_{rl}+(\mu_l+\mu_r)\big[i^{(1)}_{lr}+i^{(1)}_{rl}\big]\nonumber\\ & & -\mu_l\mu_r\big[i^{(0)}_{lr}+i^{(0)}_{rl}\big]+\hbar\omega\big[i^{(1)}_{lr}+i^{(1)}_{rl}\big]\nonumber\\ & & -\hbar\omega\mu_li^{(0)}_{lr } -\hbar\omega\mu_ri^{(0)}_{rl } \big]~.\end{aligned}\ ] ] we report the expressions of the integrals given by eqs . ( [ i0]-[i2 ] ) and
factorize the various contributions .
it gives @xmath140\nonumber\\ & & + \frac{(\hbar\omega - ev)^3}{6}n(\hbar\omega - ev)\nonumber\\ & & + \frac{(\hbar\omega+ev)^3}{6}n(\hbar\omega+ev)\bigg]~,\end{aligned}\ ] ] which reduces at equilibrium ( zero - voltage ) to @xmath141n(\hbar\omega)~,\nonumber\\\end{aligned}\ ] ] and at zero - temperature to @xmath142~,\end{aligned}\ ] ] since we have @xmath121 when @xmath122 .
note that @xmath143 is obtained from the expression of @xmath144 by inverting the voltage @xmath137 , as a consequence we have @xmath145 .
+ for @xmath86 , we have @xmath146\nonumber\\ & = & \frac{e^{2}}{h}(2\delta_{\alpha\beta}-1)[i^{(0)}_{ll}+i^{(0)}_{rr}]\nonumber\\ & = & \frac{e^{2}}{h}(2\delta_{\alpha\beta}-1)2\hbar\omega n(\hbar\omega)~.\end{aligned}\ ] ] for @xmath86 , we have @xmath147\nonumber\\ & & = \frac{e}{h}\left[i^{(1)}_{ll}+i^{(1)}_{rr}-\left(\mu_l+\frac{\hbar\omega}{2}\right)[i^{(0)}_{ll}+i^{(0)}_{rr}]\right]~.\nonumber\\\end{aligned}\ ] ] after simplification , it leads to @xmath148 a similar calculation leads to @xmath149 .
moreover , we have @xmath150 , and @xmath151 . for @xmath86
, we have @xmath152\nonumber\\ & & \times[f^e_l(\varepsilon)f^h_l(\varepsilon-\hbar\omega)+f^e_r(\varepsilon)f^h_r(\varepsilon-\hbar\omega)]\nonumber\\ & & + \frac{\hbar^2\omega^2}{4}\sum_{\gamma\delta}f^e_\gamma(\varepsilon)f^h_\delta(\varepsilon-\hbar\omega)\bigg)\nonumber\\ & & = \frac{1}{h}\bigg[i^{(2)}_{ll}+i^{(2)}_{rr}-(2\mu_l+\hbar\omega)[i^{(1)}_{ll}+i^{(1)}_{rr}]\nonumber\\ & & + \mu_l(\mu_l+\hbar\omega)[i^{(0)}_{ll}+i^{(0)}_{rr}]+\frac{\hbar^2\omega^2}{4}\sum_{\gamma\delta}i^{(0)}_{\alpha\beta}\bigg]~.\nonumber\\\end{aligned}\ ] ] it gives @xmath153~.\end{aligned}\ ] ] a similar calculation leads to @xmath154~.\end{aligned}\ ] ] in addition , we have @xmath138 and @xmath145 .
99 l. saminadayar , d.c .
glattli , y. jin , and b. etienne , phys .
lett . * 79 * , 2526 ( 1997 ) .
r. de picciotto , m. reznikov , m. heiblum , v. umansky , g. bunin , and d. mahalu , nature london * 389 * , 162 ( 1997 ) .
r. deblock , e. onac , l. gurevich , and l. kouwenhoven , science * 301 * , 203 ( 2003 ) .
g. schull , n. nel , p. johansson , and r. berndt , phys .
lett . * 102 * , 057401 ( 2009 ) .
r. snchez and m. bttiker , eur .
* 100 * , 47008 ( 2012 ) ; * 104 * , 49901 ( 2013 ) .
f. giazotto , t.t .
heikkil , a. luukanen , a.m. savin , and j.p .
pekola , rev .
phys . * 78 * , 217 ( 2006 ) .
r. snchez , b. sothmann , a.n .
jordan , and m. bttiker , new j. phys . *
15 * , 125001 ( 2013 ) .
f. battista , f .
haupt , and j. splettstoesser , j. phys .
: conf . ser . *
568 * 052008 ( 2014 ) .
a. crpieux and f. michelini , j. phys . : condens .
matter * 27 * , 015302 ( 2015 ) .
a. crpieux and f. michelini , j. stat .
mech . 054015 ( 2016 ) .
blanter and m. bttiker , phys . rep . *
336 * , 1 ( 2000 ) .
t. martin , in _ les houches session lxxxi _ , edited by h. bouchiat _ et al .
_ ( elsevier , amsterdam , 2005 ) .
r. zamoum , m. lavagna , and a. crpieux , phys .
b * 93 * , 235449 ( 2016 ) .
d. sergi , phys .
b * 83 * , 033401 ( 2011 ) . d.v . averin and j.p .
pekola , phys .
lett . * 104 * , 220601 ( 2010 ) ; europhys . lett . * 96 * , 67004 ( 2011 ) .
agarwalla , j .- h . jiang , and d. segal , phys .
b * 92 * , 245418 ( 2015 ) .
f. battista , f. haupt , and j. splettstoesser , phys .
b * 90 * , 085418 ( 2014 ) .
z. yu , g .-
tang , and j. wang , phy .
b * 93 * , 195419 ( 2016 ) .
f. zhan , s. denisov , p. hnggi , phys .
b * 84 * , 195117 ( 2011 ) ; physica status solidi ( b ) * 250 * , 2355 ( 2013 ) .
a. crpieux , f. simkovic , b. cambon , and f. michelini , phys .
b * 83 * , 153417 ( 2011 ) ; phys . rev .
b * 89 * , 239907(e ) ( 2014 ) .
a. goker and b. uyanik , physics letters a * 376 * , 2735 ( 2012 ) .
a. goker and e. gedik , j. phys . : condens .
matter * 25 * , 125301 ( 2013 ) .
a. boehnke , m. walter , n. roschewsky , t. eggebrecht , v. drewello , k. rott , m. mnzenberg , a. thomas , and g. reiss , rev .
instrum . * 84 * , 063905 ( 2013 ) .
m. bagheri tagani and h. rahimpour soleimani , int .
j. thermophys .
* 35 * , 136 ( 2014 ) .
r. chirla and c.p .
moca , phys .
b * 89 * , 045132 ( 2014 ) .
h. zhou , j. thingna , p. hnggi , j .- s .
wang , and b. li , sci .
rep . * 5 * , 14870 ( 2015 ) . a .-
dar and p. lombardo , phys .
b * 93 * , 035303 ( 2016 ) . m.f .
ludovico , f. battista , f. von oppen , and l. arrachea , phys .
b * 93 * , 075136 ( 2016 )
. h. okada and y. utsumi , arxiv:1605.03809 ( 2016 ) .
lesovik and r. loosen , pisma zh .
. fiz . * 65 * , 280 ( 1997 ) [ jetp lett .
65 , 295 ( 1997 ) ] .
u. gavish , y. levinson , and y. imry , phys .
b * 62 * , 10637 ( 2000 ) . h.j.w
. haug and a.p .
jauho , _ quantum kinetics in transport and optics of semiconductors _ , springer series in solid - state sciences 123 , second substantially revised edition ( 2010 ) . a .-
jauho , n .- s .
wingreen , and y. meir , phys .
b * 50 * , 5528 ( 1994 ) .
a. crpieux and f. michelini , int
. j. nanotechnol .
* 9 * , 355 ( 2012 ) .
j. schwinger , j. math . phys . * 2 * , 407 ( 1961 ) ; l.v .
keldysh , zh .
. fiz . * 47 * , 1515 ( 1964 ) .
j. hammer and w. belzig , phys .
b * 84 * , 085419 ( 2011 ) ; phys . rev .
b * 87 * , 125422 ( 2013 ) .
t. andrade , s.a .
gentle , and b. withers , j. high energ . phys . * 06 * , 134 ( 2016 ) .
j. basset , h. bouchiat , and r. deblock , phys .
lett . * 105 * , 166801 ( 2010 ) . for an energy independent transmission amplitude
, we have @xmath155 , and @xmath156 .
b. roussel , p. degiovanni , and i. safi , phys .
b * 93 * , 045102 ( 2016 ) . n.v .
gnezdilov , m. diez , m.j .
pacholski , and c.w.j .
beenakker , phys .
b * 94 * , 115415 ( 2016 ) .
note that at zero - temperatures , for positive voltage and in weak transmission limit , the expressions for the noises read as @xmath157/(1+p+q)!$ ] , with @xmath158 , @xmath159 , and @xmath160 , where @xmath161 is the heaviside function .
all the asymptotic values of noises given in the large temperature limit are calculated from eqs .
( [ exp_noise]-[c ] ) taking @xmath162 .
we have obtained : @xmath163 , @xmath164 , @xmath165 , and have checked numerically that @xmath166 varies linearly with temperature .
mahan , in _ many - particle physics _ ,
kluwer academic / plenum publishers ( 2000 ) . | using the keldysh green function technique , we calculate the finite - frequency correlator between the electrical current and the heat current flowing through a quantum dot connected to reservoirs . at equilibrium , we find that this quantity , called mixed noise , is linked to the thermoelectric ac - conductance by the fluctuation - dissipation theorem . out - of - equilibrium , we discuss its spectrum and find evidence of the close relationship between the mixed noise and the thermopower .
we study the spectral coherence and identify the conditions to have a strong correlation between the electrical and heat currents . the change in the spectral coherence due to the presence of a temperature gradient between the reservoirs
is also highlighted . |
many components to build a high quality natural language processing system rely on the synonym extraction .
examples include query expansion , text summarization @xcite , question answering @xcite , and paraphrase detection . although the value of synonym extraction is undisputed , manual construction of such resources is always expensive , leading to a low knowledgebase ( kb ) coverage @xcite . in the medical domain ,
this kb coverage issue is more serious , since the language use variability is exceptionally high @xcite .
in addition , the natural language content in the medical domain is also growing at an extremely high speed , making people hard to understand it , and update it in the knowledgebase in a timely manner . to construct a large scale medical synonym extraction system ,
the main challenge to address is how to build a system that can automatically combine the existing manually extracted medical knowledge with the huge amount of the knowledge buried in the unstructured text . in this paper , we construct a medical corpus containing 130 m sentences ( 20 gigabytes pure text ) .
we also construct a semi - supervised framework to generate a vector representation for each medical term in this corpus .
our framework extends the word2vec model @xcite by integrating the existing medical knowledge in the model training process . to model the concept of synonym ,
we build a `` * * concept space * * '' that contains both the semi - supervised term embedding features and the expanded features that capture the similarity of two terms on both the word embedding space and the surface form .
we then apply a linear classifier directly to this space for synonym extraction .
since both the manually extracted medical knowledge and the knowledge buried under the unstructured text have been encoded in the concept space , a cheap classifier can produce satisfying extraction results , making it possible to efficiently process a huge amount of the term pairs .
our system is designed in such a way that both the existing medical knowledge and the context in the unstructured text are used in the training process .
the system can be directly applied to the input term pairs without considering the context .
the overall contributions of this paper on medical synonym extraction are two - fold : * from the perspective of applications , we identify a number of unified medical language system ( umls ) @xcite relations that can be mapped to the synonym relation ( table [ tbl : synonymrelations ] ) , and present an automatic approach to collect a large amount of the training and test data for this application .
we also apply our model to a set of 11b medical term pairs , resulting in a new medical synonym knowledgebase with more than 3 m synonym candidates unseen in the previous medical resources . * from the perspective of methodologies , we present a semi - supervised term embedding approach that can train the vector space model using both the existing medical domain knowledge and the text data in a large corpus .
we also expand the term embedding features to form a concept space , and use it to facilitate synonym extraction .
the experimental results show that our synonym extraction models are fast and outperform the state - of - the - art approaches on medical synonym extraction by a large margin .
the resulting synonym kb can also be used as a complement to the existing knowledgebases in information extraction tasks .
a wide range of techniques has been applied to synonym detection , including the use of lexicosyntactic patterns @xcite , clustering @xcite , graph - based models @xcite@xcite@xcite@xcite and distributional semantics @xcite@xcite@xcite@xcite@xcite .
there are also efforts to improve the detection performance using multiple sources or ensemble methods @xcite@xcite@xcite .
the vector space models are directly related to synonym extraction .
some approaches use the low rank approximation idea to decompose large matrices that capture the statistical information of the corpus .
the most representative method under this category is latent semantic analysis ( lsa ) @xcite .
some new models also follow this approach like hellinger pca @xcite and glove @xcite .
neural network based representation learning has attracted a lot of attentions recently .
one of the earliest work was done in @xcite .
this idea was then applied to language modeling @xcite , which motivated a number of research projects in machine learning to construct the vector representations for natural language processing tasks @xcite@xcite@xcite@xcite@xcite@xcite . following the same neural network language modeling idea ,
word2vec @xcite significantly simplifies the previous models , and becomes one of the most efficient approach to learn word embeddings . in word2vec , there are two ways to generate the `` input - desired output '' pairs from the context : `` skipgram '' ( predicts the surrounding words given the current word ) and `` cbow '' ( predicts the current word based on the surrounding words ) , and two approaches to simplify the training : `` negative sampling '' ( the vocabulary is represented in one hot representation , and the algorithm only takes a number of randomly sampled `` negative '' examples into consideration at each training step ) and `` hierarchical softmax '' ( the vocabulary is represented as a huffman binary tree ) .
so one can train a word2vec model from the input under 4 different settings , like `` skipgram''+negative sampling " .
word2vec is the basis of our semi - supervised word embedding model , and we will discuss it with more details in section [ sec : conceptembedding ] .
our medical corpus has incorporated a set of wikipedia articles and medline abstracts ( 2013 version ) .
we also complemented these sources with around 20 medical journals and books like _ merck manual of diagnosis and therapy_. in total , the corpus contains about 130 m sentences ( about 20 g pure text ) , and about 15 m distinct terms in the vocabulary set .
a significant amount of the medical knowledge has already been stored in the unified medical language system ( umls ) @xcite , which includes medical concepts , definitions , relations , etc .
the 2012 version of the umls contains more than 2.7 million concepts from over 160 source vocabularies .
each concept is associated with a unique keyword called cui ( concept unique identifier ) , and each cui is associated with a term called preferred name .
the umls consists of a set of 133 subject categories , or semantic types , that provide a consistent categorization of all cuis .
the semantic types can be further grouped into 15 semantic groups .
these semantic groups provide a partition of the umls metathesaurus for 99.5% of the concepts .
domain specific parsers are required to accurately process the medical text .
the most well - known parsers in this area include metamap ( aronson , 2001 ) and medicalesg , an adaptation of the english slot grammar parser @xcite to the medical domain .
these tools can detect medical entity mentions in a given sentence , and automatically associate each term with a number of cuis .
not all the cuis are actively used in the medical text .
for example , only 150k cuis have been identified by medicalesg in our corpus , even though there are in total 2.7 m cuis in umls .
synonymy is a semantic relation between two terms with very similar meaning .
however , it is extremely rare that two terms have the exact same meaning . in this paper ,
our focus is to identify the near - synonyms , i.e. two terms are interchangeable in some contexts @xcite .
the umls 2012 release contains more than 600 relations and 50 m relation instances under 15 categories .
each category covers a number of relations , and each relation has a certain number of cui pairs that are known to bear that relation . from umls relations , we manually choose a subset of them that are directly related to synonyms , and summarize them in table [ tbl : synonymrelations ] . in table
[ tbl : synonymexamples ] , we list several synonym examples provided by these relations .
.umls relations corresponding to the synonym relation , where `` ro '' stands for `` has relationship other than synonymous , narrower , or broader '' , `` rq '' stands for `` related and possibly synonymous '' , and `` sy '' stands for `` source asserted synonymy '' .
[ cols="^,^",options="header " , ] our method was very scalable .
it took on average several hours to generate the word embedding file from our medical corpus with 20 g text using @xmath0 g cpus and roughly 30 minutes to finish the training process using one cpu . to measure the scalability at the apply time
, we constructed a new medical synonym knowledgebase with our best synonym extractor .
this was done by applying the concept space model trained under the neg+skip setting to a set of 11b pairs of terms .
all these terms are associated with cuis , and occur at least twice in our medical corpus .
this kb construction process finished in less than 10 hours using one cpu , resulting in more than 3 m medical synonym term pairs . to evaluate the recall of this knowledgebase
, we checked each term pair in the held out synonym dataset against this kb , and found that more than @xmath1 of them were covered by this new kb .
precision evaluation of this kb requires a lot of manual annotation effort , and will be included in our future work . in section [ sec : featureexpansion ] , we expand the raw features with the matching features and several other feature expansions to model the term relationships . in this section , we study the contribution of each individual feature to the final results .
we added all those expanded features to the raw features one by one and re - ran the experiments for the concept space model .
the results and the feature contributions are summarized in table [ tbl : featurecontributions ] .
the results show that adding matching features , `` sum '' and `` difference '' features can significantly improve the @xmath2 scores .
we can also see that adding the last two feature sets does not seem to contribute a lot to the average @xmath2 score .
however , they do contribute significantly to our best @xmath2 score by about @xmath3 .
in this paper , we present an approach to construct a medical concept space from manually extracted medical knowledge and a large corpus with 20 g unstructured text .
our approach extends the word2vec model by making use of the medical knowledge as extra label information during the training process .
this new approach fits well for the medical domain , where the language use variability is exceptionally high and the existing knowledge is also abundant .
experiment results show that the proposed model outperforms the baseline approaches by a large margin on a dataset with more than one million term pairs .
future work includes doing a precision analysis of the resulting synonym knowledgebase , and exploring how deep learning models can be combined with our concept space model for better synonym extraction .
r. collobert and j. weston . a unified architecture for natural language processing : deep neural networks with multitask learning . in _ proceedings of the 25th international conference on machine learning _ , 2008 .
x. glorot , a. bordes , and y. bengio .
domain adaptation for large - scale sentiment classification : a deep learning approach . in _ proceedings of the 28th international conference on machine learning _ , 2011 .
a. henriksson , m. conway , m. duneld , and w. w. chapman .
identifying synonymy between snomed clinical terms of varying length using distributional analysis of electronic health records . in _ proceedings of amia annual symposium .
_ , 2013 .
a. henriksson , m. skeppstedt , m. kvist , m. conway , and m. duneld .
corpus - driven terminology development : populating swedish snomed ct with synonyms extracted from electronic health records . in _ proceedings of bionlp _ , 2013 .
p. huang , x. he , j. gao , l. deng , a. acero , and l. heck .
learning deep structured semantic models for web search using clickthrough data . in _ proceedings of the acm international conference on information and knowledge management ( cikm ) _ , 2013 .
t. mikolov , k. chen , g. corrado , and j. dean .
efficient estimation of word representations in vector space . in _ proceedings of the workshop at international conference on learning representations _ , 2013 .
y. peirsman and d. geeraerts .
predicting strong associations on the basis of corpus data . in _ proceedings of the 12th conference of the european chapter of the association for computational linguistics _ , 2009 .
r. socher , c. c. lin , a. ng , and c. d. manning .
parsing natural scenes and natural language with recursive neural networks . in _ proceedings of the 28th international conference on machine learning ( icml ) _ , 2011 . | in this paper , we present a novel approach for medical synonym extraction .
we aim to integrate the term embedding with the medical domain knowledge for healthcare applications .
one advantage of our method is that it is very scalable .
experiments on a dataset with more than 1 m term pairs show that the proposed approach outperforms the baseline approaches by a large margin . |
a landmark in the study of extragalactic radio sources was the demonstration by fanaroff and riley ( @xcite ) of the existence of a relatively sharp morphological transition at a radio luminosity of @xmath0 corresponding to @xmath1 w hz@xmath2 .
the great majority of sources below this luminosity ( fr i type ) are characterized by having diffuse radio lobes , with their brightest regions within the inner half of the radio source , and so can be called edge - dimmed . on the contrary ,
more powerful sources are usually straighter , exhibit edge - brightened ( fr ii ) morphology , and typically contain hotspots near the outer edges of their radio lobes .
more recently it was realized that the critical radio luminosity separating the fri and fr ii actually increases with the optical luminosity of the host elliptical galaxy , so that @xmath3 ( as measured from fig . 1 of ledlow & owen @xcite ) ; also see owen & white ( @xcite ) ; owen & ledlow ( @xcite ) for earlier indications of this effect . although many detailed differences between the host galaxies of the fr i and fr ii sources have been discovered ( e.g. , baum et al .
@xcite ; zirbel @xcite ; for a recent summary , see gopal - krishna & wiita @xcite ) , it is fair to state that a key fact is that the more luminous the host galaxy is , the more powerful the radio source must be in order to attain the fr ii morphological classification .
the most widely known explanation for the difference between the fr i and fr ii sources is that , while the jets in both cases start out moving at very high ( relativistic ) speeds , those in fr ii sources remain that way out to multi - kpc distances , while those in the fr i s decelerate to much lower speeds within a few kpc of the core ( e.g. , begelman @xcite ; bicknell @xcite ; de young @xcite ; komissarov @xcite ; laing @xcite , @xcite ; laing et al .
detailed models for decelerating relativistic jets were developed by bicknell ( @xcite , @xcite ) .
the core of his argument is that relativistic jets , once they come into pressure equilibrium with the external interstellar medium ( ism ) or intracluster medium ( icm ) , will become strongly kelvin - helmholtz unstable and then entrain substantial amounts of cold ambient material .
bicknell ( @xcite ) showed that this instability typically set in as the internal jet mach number approaches @xmath4 , corresponding to internal bulk velocities having dropped to about @xmath5 . of course , many other explanations for the fr
i / fr ii dichotomy have been proposed ( e.g. , baum et al .
@xcite ; reynolds et al .
@xcite , @xcite ; meier @xcite ; valtonen & heinmki @xcite ; see the recent discussion by gopal - krishna & wiita @xcite ) .
some of these proposals involve fundamental differences in the nature of the jets ( electron / positron vs. electron / proton plasma ) or of the central engine ( black hole spin , type of accretion disk ) .
the viability of any member of this class of explanations is challenged by the existence of hybrid morphology radio sources , or hymors , which have a clear fr i morphology on one side of the host galaxy , but a distinct fr ii morphology on the other side ( gopal - krishna & wiita ( @xcite , @xcite ) .
an alternative approach posits that fr i and fr ii sources differ primarily in the importance of the beam thrust relative to the basic parameters of the ambient medium ( gopal - krishna & wiita @xcite ( gkw88 ) ; gopal - krishna @xcite ( gk91 ) ; gopal - krishna et al .
@xcite ( gkwh96 ) ; blandford 1996 ) . in this version of the deceleration scenario ,
the emphasis is on the slowing of the advance of the hotspot , or working surface , at the end of the jet , rather than on the slowing of the bulk flow within the jet ( e.g. , bicknell @xcite ) .
when the hotspot s advance becomes transonic relative to the ambient medium , its mach disk weakens considerably due to the fall in ram pressure , and the jet becomes decollimated ; this soon leads to an fr i morphology .
further expansion of such jets beyond this point is expected to be in the form of a plume , as discussed in gkw88 , where it was also shown that for reasonable values of jet and galaxy parameters ( density , core radius and temperature of the ism ) such a jet flaring can occur within several kpc of the core for jet powers below about @xmath6 erg s@xmath2 , consistent with the observations of the fr i / fr ii division ( gkw88 ; gk91 ; gkwh96 ) .
also , in this picture , the concomitant dimunition / cessation of the ` backflow ' of the beam plasma would cause a depletion of the protective sheath of beam plasma around the jet , facilitating entrainment of the ism material into the jet flow .
bicknell ( @xcite ) expanded his model to account for the owen
ledlow transition in the @xmath7@xmath8 plane .
a key ingredient in his scenario involved connecting the jet dynamics to empirically established relations between the optical magnitude of the galaxy on the one hand , and the soft x - ray luminosity , core radius , and central velocity dispersion of elliptical galaxies on the other hand . in the present study , we follow bicknell and adopt these same empirical relationships .
bicknell then ties the galactic parameters to those of the jet by demanding that the establishment of a pressure balance between the jet and the external medium corresponds closely to the location of the jet s internal transonic transition . by doing so
, he was able to derive a formula connecting beam power to the optical luminosity of the host galaxy . in this paper
we follow a different approach in this latter stage .
additional assumptions on the efficiency of the conversion of the jet s energy into total radio luminosity ( @xmath9 ) , and thence into monochromatic radio emission , allowed bicknell ( @xcite ) to obtain a fairly good fit to the slope of the owen ledlow division , finding that @xmath10 , as well as getting a decent fit to the intercept ; however , there were quite a few poorly constrained parameters in his model .
nevertheless , bicknell ( @xcite ) argued that the the slope of the relation was rather insensitive to the likely uncertainties in parameters .
he further argued that the intercept would tend to be driven towards better agreement with the data for plausible variations in those parameters , which included : the ratio of the jet s lifetime to the synchrotron cooling time of the highest energy electrons ; the ratio of the upper and lower cutoff energies for the electron distribution ; the ratio of jet radius to its length at the transition distance ; the ratio of the product of the jet pressure and square of its radius evaluated at the transition radius to that of the external medium evaluated at its core radius ; the bulk velocity of the jet at the transition point ; the low frequency radio spectral index ; and the value of an integral ( discussed below ) which depends on an upper cut - off radius for the x - ray emitting halo of the host galaxy .
the new magnetic switch model ( meier @xcite ) fundamentally distinguishes fr i from fr ii sources through the different speeds of rotation of the magnetic field lines of their central engines , which are tied to the different spin rates of their supermassive black holes . while this scenario can also produce a slope for the @xmath11 dividing line close to that found by ledlow & owen ( @xcite ) ( meier @xcite ) we recall that it is hard to reconcile this scheme with the existence of hymors ( gopal - krishna & wiita @xcite , @xcite ) .
venturing beyond the standard nuclear jet paradigm , gravitational slingshot models can in principle be compatible with the owen
ledlow relation ( valtonen & heinmki @xcite ) , and could also account for hymors ( gopal - krishna & wiita @xcite ) .
in an earlier study involving `` weak headed quasars '' , where prominent one - sided jets are not seen to terminate in a conspicuous hotspot , we have argued that dissipation of jet power , as suggested by swarup et al .
( @xcite ) and saikia et al . ( @xcite ) , may actually not be responsible for the lack of terminal hotspots ( gkwh96 ) . taking a clue from the fact that no two - sided jets are seen in these sources
, we argued that the lack of a hotspot could be best explained through the onset of a jet s decollimation when the hotspot s ( or , nearly equivalently , the bow shock s ) velocity becomes transonic relative to the external medium ( gkwh96 ) .
this successful explanation of this class of radio sources motivates us to seek an explanation for the owen
ledlow relation in terms of a similar scenario .
this model makes no assumptions about the relativistic nature of the bulk velocity of the internal jet fluid , which may be gradually decelerating as the head of the jet advances .
our study makes use of the same empirical relations between the elliptical s blue magnitude , @xmath12 , and its soft x - ray emission , @xmath13 ( donnelly , faber and oconnell @xcite ) , stellar velocity dispersion , @xmath14 ( terlevich et al
. @xcite the faber jackson relation ) , and x - ray core radius , @xmath15 ( kormendy @xcite ) , as employed by bicknell ( @xcite ) , assuming @xmath16
km s@xmath2 mpc@xmath2 : @xmath17 @xmath18 @xmath19 however , we adopt the initially conical jet model based on our earlier work ( gopal - krishna & wiita @xcite ) , which was generalized in gopal - krishna et al .
( @xcite , gkws89 ) to allow for bulk relativistic jet flow whereby the hotspot motion can also be relativistic .
ram pressure balance , under these conditions , gives the hotspot velocity , @xmath20 , as a function of distance , @xmath21 , from the central engine , as ( gkws89 ) : @xmath22^{\delta/2 } } { d + x[1+(d / a)^2]^{\delta/2}}}.\ ] ] here @xmath23 , where @xmath24 is the jet ( beam ) power , @xmath25 is the jet s effective opening angle , @xmath26 and @xmath27 have their usual definitions , and @xmath28 is the central density of the ism , which is taken to fall off as ( e.g. , forman et al .
@xcite ; canizares et al .
@xcite ; conway @xcite ) @xmath29^{\delta}}}.\ ] ] for our calculations we choose @xmath30 rad , which is certainly justified as appropriate for the inner jet regions , particularly since it is now recognized that the jet s thrust acts on a larger area than the instantaneous hot spot ( e.g. , scheuer s @xcite ` dentist - drill ' , as supported by many three - dimensional numerical simulations : norman @xcite ; clarke @xcite ; kaiser & alexander @xcite ; hooda & wiita @xcite ) . following bicknell ( @xcite ) , we further assume that the x - ray emitting ism gas temperature is tied to the central stellar velocity dispersion , @xmath14 , via , @xmath31 , which is fall s ( @xcite ) relation between circular velocity and @xmath14 . here
@xmath32 is roughly 0.75 , as inferred from soft x - ray images of nearby ellipticals ( e.g. , forman et al .
@xcite ; canizares et al .
@xcite ; sarazin @xcite ) , as well as dynamical models of radio sources ( e.g. , gopal - krishna & wiita @xcite ; conway @xcite ) .
studies of many sub - classes of young radio sources , including compact symmetric objects ( e.g. , owsianik & conway @xcite ; conway @xcite ) , gigahertz peaked sources ( carvalho @xcite ) and compact steep spectrum radio sources ( gopal - krishna & wiita @xcite ; jeyakumar et al .
@xcite ) have indicated that the mean density of the gas interacting with the radio lobes in the inner @xmath33 kpc or so is typically a few atoms @xmath34 , or several times that estimated from x - ray emission .
this evidence in favor of a significant contribution of cooler gas to the ambient medium ( at least in the inner portions of elliptical galaxies ) is further supported by studies of the linear - size distribution of these radio sources ( odea & baum @xcite ) .
we shall parameterize the relation between the confining central density , @xmath28 , used in eq .
( 5 ) and that derived from x - ray measurements , @xmath35 as follows : @xmath36 , with @xmath37 ; the fiducial value we use below is @xmath38 . the radio data cited above as well as the relation between @xmath39 and the x - ray core radius ( eq . 3 ) lead to values of of @xmath40 kpc .
therefore the ism density begins to approach the large scale behavior , and thus declines quite rapidly , at radial distances beyond a few kpc .
hence , the most likely regime for the jets decollimation due to the hot spots having slowed to subsonic speeds lies within roughly 10 kpc of the core .
if the hot spot manages to retain supersonic speed out to such radial distances , then in most cases the jet is likely to propagate ahead down the rapidly declining ism density , preserving its fr ii character . in some cases
another possibility for jet flaring can arise farther out when the jet crosses the pressure - matched interface between the ism and the intergalactic medium ( gkw88 ; wiita et al .
@xcite ; wiita & norman @xcite ; hooda et al .
@xcite ; hooda & wiita @xcite , @xcite ; zhang et al .
relatively few sources will , however , make a transition to fr i morphology upon crossing this interface since a significant degree of beam collimation appears to occur by such radial distances within the extended halos of most galaxies ( blundell et al .
@xcite ; jeyakumar & saikia @xcite ) . for such very extended sources ,
our model is clearly oversimplified , as the evolution appears to break from self - similarity and the decrease in @xmath7 with time ( even for constant @xmath24 ) can not be ignored ( e.g. , blundell & rawlings @xcite ) .
other types of sources that may not be readily accommodated in our simple picture are , for instance , cases where a narrow jet is seen to terminate in a bright knot far away ( say @xmath41 kpc ) from the nucleus , and flares only beyond that point ( e.g. , 3c 130 , hardcastle 1998 ) .
of course , higher values of @xmath42 are expected for sources for which the beam s power and opening angle lie near the upper and lower ends of their respective ranges .
also , the possibility of a bright knot and jet expansion occurring when the jet interacts with irregularities in the ambient medium must be considered ( e.g. , wang et al .
@xcite ) . in this context
it is especially interesting to recall that recent radio observations of the nearest radio galaxy , centaurus a , have provided clear evidence that a radio jet can even emerge recollimated after flaring upon hitting a gas cloud ( morganti et al .
these australian telescope compact array observations have revealed that the flaring of the northern jet of cen a , a few kiloparsecs away from from the nucleus , gives rise to the inner radio ` lobe ' , out of which a narrow jet is seen to emerge and propagate further out for another few kiloparsecs before flaring and fading away in the fashion of typical fr i jets .
morganti et al .
( @xcite ) have interpreted this striking ` leaky bulb ' type morphology as being a manifestation of the de laval nozzle model proposed by blandford & rees ( @xcite ) . in our analytical formulation ,
we evaluate the critical beam power , @xmath43 , for which the hot - spot deceleration to subsonic velocities occurs at a fiducial distances , @xmath42 , of 3 and 10 kpc from the core , for a range of absolute magnitudes of the host galaxy between @xmath44 and @xmath45 .
the 10 kpc value is a typical distance at which jets flare in a sample of radio galaxies ( odonoghue et al .
@xcite ) . a recent study of a sample of 38 fr i sources by laing et al .
( @xcite ) gives a mean projected value of @xmath46 kpc for the radial distance of the point where the kiloparsec scale jet first becomes visible , after an initial emission gap ( using our value of @xmath47 ) .
our main results are not very sensitive to the choice of @xmath42 , at least until the very most luminous galaxies are considered , and the predicted slope of the separation between fr i and fr ii sources that we derive below is not strongly dependent upon it .
we now write the relevant expressions as follows : @xmath48 where @xmath49 for @xmath50 .
we also require , @xmath51 , where the cooling function , @xmath52 is a rather weak function of the temperature ; assuming @xmath53k , @xmath54 ( bicknell @xcite ) , and the integral over the gas density distribution gives @xmath55 ( for @xmath56 ) , a value we adopt , following bicknell ( @xcite ) . with these choices , @xmath57 equations ( 13 ) and ( 6 )
can now be combined to yield : @xmath58 @xmath59 @xmath60
we next impose the transonic condition on the lhs of eq .
( 4 ) , via @xmath61^{1/2 } \sigma.\ ] ] taking the ratio of specific heats for the combined plasma , @xmath62 , and using @xmath56 , this becomes @xmath63 equating this to the rhs of ( 4 ) , treated as a function of @xmath64 , gives , @xmath65^{\delta/2 } + x}},\ ] ] as the relation that must be solved for @xmath43 in terms of @xmath39 .
the values of @xmath43 , @xmath15 , and @xmath28 can be found for specified values of @xmath39 and @xmath42 by solving eqs .
( 3 ) , ( 8) , ( 9 ) and ( 12 ) .
the results are given in table 1 , along with , @xmath66 , the local slope of the @xmath67@xmath68 relation .
for the choice of @xmath69 kpc , these slopes , @xmath66 , vary little for @xmath70 , and correspond closely to a relation where @xmath71 ; for @xmath72 kpc , @xmath66 is only slightly less steep .
for the most luminous galaxies , the value of @xmath66 from the model flattens somewhat . in the absence of a definitive approach to compute the efficiency , @xmath73 , of conversion of the beam power into synchrotron radiation , we make the commonly adopted first order approximation that the monochromatic radio luminosity is a constant fraction of the beam power ( at least near the transition luminosity ) . then the local slope remains the same , with @xmath74 , and the key features of the owen ledlow diagram are satisfactorily explained , with a very good fit to the slope retained .
the intercept also fitted well , provided the product @xmath75 for @xmath69 kpc . as shown in fig . 1
, this is in excellent agreement with the bifurcation between fr i and fr ii radio sources , and nominally provides even a better fit than does bicknell s ( @xcite ) model , which yields a power - law index above 2.0 .
we use a mean color index for elliptical galaxies @xmath76 taken from a ccd photometric study of cluster galaxies ( jrgensen et al .
@xcite ) to convert from the absolute blue magnitudes used in eqs . ( 13 ) to the red magnitudes plotted by ledlow et al .
( @xcite ) ; the dependence of color index on optical magnitude is small over this range and we have not attempted to correct for it ( jrgensen et al .
@xcite ; kodama et al .
@xcite ; peletier et al .
@xcite ) .
we note that earlier estimates based on the application of beam models for fr ii sources are in reasonable agreement with such values of @xmath77 ( gopal - krishna & saripalli @xcite ; dreher @xcite ; saripalli & gopal - krishna @xcite ) , particularly if a `` ready to radiate '' electron - positron jet composition is assumed ; and evidence for such a composition is growing ( e.g. , reynolds et al .
@xcite ; kaiser & alexander @xcite ; kaiser et al .
@xcite ; hirotani et al .
@xcite ) . given that our analysis involves a fairly large set of empirical relations for elliptical galaxies , the agreement of our prediction with the observational data is quite encouraging ( fig . 1 ) .
it is perhaps worth remarking on the consistency of our model with the recent very long baseline interferometry measurements for the expansion of very small and young compact symmetric object ( cso ) radio sources ( e.g. , owsianik & conway @xcite ( oc ) ) . for the cso 0710@xmath78439 ,
the best values obtained by oc are @xmath79 at a distance @xmath80pc ( @xmath81 ) , for @xmath82 @xmath34 , and @xmath83erg s@xmath2 . inserting the last three values into eq .
( 4 ) , and choosing our nominal @xmath30 rad , we obtain @xmath84 , in quite reasonable agreement ; a somewhat higher value for @xmath25 might be inferred from the maps in oc , thereby reducing our computed velocity and improving the agreement .
less well determined values of expansion speeds obtained by other groups for 3 other cso radio sources quoted by oc range from @xmath85 and would easily be accommodated by more typical values of @xmath24 .
the derived range , @xmath86 , for 0710@xmath78439 ( oc ) is also in good agreement with our model .
the forgoing analysis is admittedly oversimplified ; for instance , we have considered a fixed opening angle for beams of all powers , and have also assumed specific values of @xmath32 and @xmath87 , all of which are certain to vary somewhat from source to source . on the other hand ,
the division between fr i and fr ii sources is not perfect by any means , and such a spread in properties would allow for the small number of sources found on the ` wrong ' side of the dividing line in fig . 1 .
also , we have ignored any variation in the jet disruption length with the galaxy s optical luminosity by typically choosing @xmath88 10 kpc ; however , such a variation is quite plausible , and therefore we have also provided results for @xmath88 3 kpc in the last two columns in table 1 . to obtain essentially as good a fit with @xmath88 3 kpc , one would have to raise @xmath73 to 0.16 if @xmath38 instead of @xmath89 for @xmath88 10 kpc ; of course both values could drop if @xmath90 is considered . as a first step towards a more realistic model , one could consider a distribution of 210 kpc for @xmath42 over the range in the hosts optical magnitudes , taking smaller @xmath42 for lower @xmath8 .
then the predicted slope of the @xmath11 relation clearly would be slightly steeper ( cf .
columns 4 and 6 of table 1 ) , further improving the accord with the data presented by ledlow & owen ( @xcite ) and recently updated by ledlow et al .
( @xcite ) .
we also note that the exact results quoted assume a specific spectral index ( @xmath91 ) and upper and lower cut - offs to the radio band ( 10 mhz and 100 ghz ) respectively , in relating the monochromatic radio power to the beam power . assuming a flatter overall spectrum , say @xmath92 , would not affect the slope of the relation but would demand that the product @xmath93 rise to @xmath94 ( for @xmath88 10 kpc ) to normalize it . despite these uncertainties , it is probably fair to note that overall our variant model employs several fewer parameters than does bicknell s ( @xcite ) , although in connecting optical and x - ray properties we have closely followed his approach .
the available very detailed maps of a few kiloparsec - scale fr i jets provide evidence for the existence of a slower moving sheath of synchrotron plasma surrounding a relativistic spine ( e.g. , laing @xcite , @xcite ; laing et al .
@xcite ) . in bicknell s picture
this sheath arises from turbulent mixing of the ambient plasma after the jet s mach number becomes low enough to allow an exponential growth of the planar kelvin - helmholtz instability . in our picture
, the entrainment process is accelerated once the hotspot s motion has become subsonic , as this results in the cessation of the ` backflow ' of the jet plasma ; this leads to the diminution of the protective cocoon around the jet core which had hitherto separated the latter from the ism material .
we also recall that the impingement of the ` backflow ' onto the jet engenders some of the reconfinement shocks ( e.g. norman et al .
@xcite ; hooda & wiita @xcite , @xcite ) so that the depletion of the shrouding cocoon plasma reduces their effects and therefore contributes to the jet s decollimation .
thus , in our picture , the dramatic slowing down of the beam flow can be thought of as an eventual outcome of the decollimation of the beam s head ( leading to fr i structure ) , instead of being its principal cause .
we thank michael ledlow and frazer owen for providing the data used in fig . 1 and b. premkumar for assistance in producing that figure .
gk appreciates hospitality at princeton university while much of this work was carried out .
pjw acknowledges support from cst funds at princeton and rpe funds at georgia state university . | we show that a model for radio source dynamics we had earlier proposed can readily reproduce the relationship between the radio power division separating the two fanaroff - riley classes of extragalactic radio sources and the optical luminosity of the host galaxy , as found by owen and ledlow ( 1994 ) . in our scenario ,
when less powerful jets eventually slow down to the point that the advance of the working surface ( i.e. , hotspot ) becomes subsonic with respect to the external gas , the jet s collimation is severely weakened .
this criterion distinguishes the powerful and well collimated fr ii sources from the weaker sources producing the less collimated fr i type morphologies . |
quantum dots fabricated within carbon nanotubes ( cnts ) have attracted considerable attention in recent years ( for reviews see _ e.g. _ [ ] ) . such devices , in which electrons are trapped within a small , strongly interacting region of the cnt by an applied electric field , show remarkable electronic transport properties@xcite and may have useful applications in future technology.@xcite cnt - based devices are of particular interest due to their doubly - degenerate orbital structure which , combined with electron spin , generates a wealth of basic physics .
one such phenomenon is the kondo effect,@xcite resulting from strong electron interactions within the dot .
it is observed@xcite when the device is tuned so that the dot has a partially - filled shell of electrons : on lowering the temperature , the dot s spin / orbital degrees of freedom become strongly coupled@xcite to those of the leads , leading to a complex many - body ground state with an enhanced electronic conductance.@xcite understanding this effect is particularly important for cnt dots , because the involvement of both spin and orbital degrees of freedom generates an @xmath0 kondo effect that persists to considerably higher temperatures up to a few kelvin ( _ i.e. _ a few tenths of an @xmath2 ) than the standard @xmath1 kondo effect in semiconductor devices.@xcite using a range of many - body techniques , the theory of the @xmath0 kondo effect is now well established@xcite and _ e.g. _ has been shown@xcite to be in good agreement with experiments@xcite performed in the absence of a magnetic field .
another consequence of the interplay between spin and orbital degrees of freedom in cnt devices is spin - orbit ( so ) coupling .
its existence was beautifully demonstrated in experiments@xcite on a very strongly correlated cnt dot , where it generates a splitting of sequential tunneling spectra at finite bias , and kinks in the magnetic field dependence of the coulomb blockade ` staircase ' at zero bias .
the strength of the so coupling was measured@xcite , and for the device studied found to be of order @xmath3@xmath4 , varying somewhat between different electron shells . comparing the typical energy scales of the kondo effect and the so coupling ,
it is striking that _ both _ may arise on the energy scale of a few tenths of an @xmath2 .
two related questions then arise : what effect does so coupling have on the standard @xmath0 kondo theory , and is the kondo / so competition seen experimentally ?
we seek to answer these questions in this paper .
aspects of spin - orbit coupling in cnts have recently been considered theoretically .
the origins of the coupling itself have been determined from direct microscopic calculation,@xcite showing that while a number of distinct spin - orbit interactions arise in principle , the dominant contribution is the direct coupling between each electron s spin and orbital angular momentum .
the effect of this coupling on the states of the isolated dot has been analysed in detail,@xcite and the resulting sequential tunneling transport properties ( arising when the kondo scale is too small to be seen experimentally@xcite ) have been calculated and compared to experiment.@xcite aspects of the competition between spin - orbit coupling and the kondo effect have also been studied@xcite via an equation of motion decoupling scheme in the @xmath5 limit . within this
rather crude approximation,@xcite the splitting of the @xmath0 kondo resonance was examined for finite so coupling and magnetic field , and an orbital kondo effect found at finite field in the two - electron coulomb blockade valley.@xcite in the present work we consider the two - fold orbitally degenerate @xmath0 anderson impurity model ( aim ) in a magnetic field , with so coupling , and study it using the numerical renormalization group ( nrg ) @xcite backed up by simple physical arguments .
nrg is ideally suited to the problem , being known for similar quantum impurity models to provide numerically exact results on the low - energy / temperature scales relevant to experiment .
the model itself also has a strong track - record , a previous nrg study@xcite of the @xmath0 aim in the absence of a magnetic field having shown that the all - important low - energy kondo physics is well reproduced when the bare model parameters are fitted to high - energy conductance features such as the coulomb blockade diamonds .
the paper is laid out as follows .
the model is described in sec .
[ sec : model ] , and the relevant theoretical background and nrg technique are discussed in sec .
[ sec : theory ] .
the behavior of the model in the atomic ( lead - uncoupled dot ) limit is outlined in sec . [ sec : atomic ] , from which simple arguments are then used to deduce the effect of introducing a finite spin - orbit interaction .
the main body of the paper is sec .
[ sec : results ] , where we present and discuss the results of nrg calculations .
we begin by considering the zero - bias conductance , as a function of gate voltage , temperature and magnetic field .
here we make comparison to the experiments of jarillo - herrero _
et al._,@xcite showing that the orbital splitting identified empirically in experiment is readily explained by the inclusion of spin - orbit coupling in the model .
we then turn to a discussion of the finite - bias conductance , comparing explicitly to the experiments of makarowski _
et al._@xcite and showing that the asymmetry observed in the kondo peaks at finite bias is also well - described by the theory .
the paper concludes with a brief summary , and a discussion of the applicability of the pure @xmath0 aim to cnt quantum dots .
the basic model used to describe a cnt quantum dot is the @xmath0 anderson impurity model@xcite , given in conventional notation by @xcite @xmath6 the ` flavor ' index @xmath7 takes four discrete values , which @xmath0 symmetry reflects physically a combination of degenerate spin and orbital degrees of freedom : @xmath8 , with @xmath9 denoting clockwise and anticlockwise orbits along the @xmath10-direction ( major axis ) of the cnt , and @xmath11 for the @xmath10-components of electron spin .
the final two terms in eqn .
( [ eq : su4ham ] ) represent the isolated dot , with orbital energy @xmath12 and charging energy @xmath13 ( @xmath14 is the dot capacitance ) ; where @xmath15 , @xmath16 is the total dot number operator , and @xmath17 creates a @xmath18-spin electron in orbital @xmath19 .
the first pair of terms describe the non - interacting conduction band ( lead ) , and tunnel coupling between the dot / lead .
each is taken to be spin and orbital conserving @xcite , reflecting physically the fact ( see _ e.g. _ [ ] ) that in clean cnt devices the leads are formed within the nanotube and so ` carry ' the orbital symmetry , which is then conserved in the tunneling process .
the model represents the experimentally relevant situation in which the single - particle level spacing of the dot exceeds both the tunnel coupling to the leads and the intradot interactions.@xcite in this case the four - electron ` shells ' of the dot are filled sequentially on sweeping the gate voltage @xmath20 ( @xmath21 ) and , for sufficiently - low temperatures and source - drain biases , only a single shell need be considered at a time . only the direct coulomb repulsion between dot electrons is moreover included ; exchange interactions are generally weaker@xcite and are not necessary@xcite to account for the experimental results of _ e.g. _ [ ] . to the ` standard model ' above
, we add the coupling to the external magnetic field , as well as the key so coupling of interest here . for a field @xmath22 applied parallel to the nanotube axis
, the zeeman coupling to the spin and orbital degrees of freedom takes the form@xcite @xmath23 where @xmath24 and @xmath25 are the @xmath10-components of the spin and orbital - pseudospin operators for orbital @xmath19 ; and where the coupling constants are @xmath26 ( with @xmath27 the electron @xmath28-factor ) and @xmath29 ( with @xmath30 the orbital moment ) .
the so interaction obtained from detailed microscopic calculations is rather complicated@xcite , but in practice only direct coupling between electron spin and the @xmath10-component of orbital angular momentum is relevant.@xcite as explained in [ ] , the so interaction is then given by [ eq : hso ] @xmath31 with @xmath32 parametrising the strength of the so coupling ; and which simple form _
e.g. _ explains in essence fully the sequential tunneling experiments of [ ] .
the full hamiltonain considered is thus @xmath33 inclusion of @xmath34 lowering the symmetry from @xmath0 to @xmath35 ( such that @xmath36 commutes with the four charge operators @xmath37 ) . finally , to connect to experiment we simply treat @xmath38 , @xmath32 and the ratio @xmath39 as parameters chosen to fit experiment ( which in practice is quite straightforward and unambiguous ) .
the dot level energy @xmath12 is proportional to the experimental gate voltage , of form @xmath40 ; but in practice it is more convenient to work with a dimensionless gate voltage@xcite @xmath41 such that differences in @xmath20 are proportional to differences in @xmath42 ( with a proportionality constant determined if desired by fit to experiment ) .
the dot green function @xmath43 is central to understanding transport through the dot ; which is directly related to the single - particle spectrum @xmath44 _ via _ the meir - wingreen formula @xcite , as now briefly summarised .
one first partitions the conduction band of eqn .
( [ eq : su4ham ] ) into two equivalent leads , left ( @xmath45 ) and right ( @xmath46 ) .
these are taken conventionally @xcite to be flat bands of width @xmath47 , with density of states @xmath48 for @xmath49 ( and with @xmath50 by far the largest energy scale in the model ) .
the tunneling matrix elements to the @xmath51 leads are taken for simplicity as @xmath52 and @xmath53 , independent of @xmath54 .
the leads are fixed at different chemical potentials , @xmath55 and @xmath56 , with a bias voltage @xmath57 between them , @xmath58 .
after transients have subsided , the bias voltage gives rise to a steady state current through the dot , @xmath59 , carried by its four conduction channels .
an exact expression for @xmath59 follows from the keldysh formalism:@xcite @xmath60 \sum_{i,\sigma } d_{i\sigma}(\omega)\ ] ] where @xmath61 @xmath62 is the hybridization strength of the dot to lead @xmath63 ( @xmath64 or @xmath46 ) , and @xmath65^{-1}$ ] is the fermi function for the lead with inverse temperature @xmath66 ( @xmath67 ) .
it is convenient to define @xmath68 ( we later take @xmath69 to be the ` unit ' of energy ) , such that the relative strength of coupling to the @xmath45 and @xmath46 leads enter through the dimensionless @xmath70 ( which can be chosen according to the experimental setup under consideration ) . in the perfectly symmetric case of @xmath71 , @xmath72 is maximal , while in the extreme asymmetric case of @xmath73 ( say ) , @xmath74 .
the key experimental quantity is the differential conductance , @xmath75 . in the zero - bias limit , gives an exact expression for @xmath76 in terms of the equilibrium single - particle spectrum : @xmath77 with @xmath78 , which further reduces to @xmath79 for @xmath80 .
@xmath81 can be calculated accurately at equilibrium using the recent fdm nrg method @xcite . to make connection to experiments at finite source - drain bias , note that while eqn .
( [ eq : j ] ) itself remains exact for finite @xmath82 , the difficulty lies in calculating @xmath81 out of equilibrium . while recent progress has been made in applying nrg to the non - equilibrium single - impurity anderson model ( see _ e.g. _ [ ] ) it is currently prohibitive to apply these techniques to the model of eqn .
( [ eq : fullh ] ) .
as in previous work @xcite we thus make the standard approximation of neglecting the @xmath57-dependence of the dot self - energy . using eqn .
( [ eq : j ] ) this leads to @xmath83 where we have taken a symmetric voltage split @xcite between the leads , @xmath84 .
equations ( [ eq : gczero])([eq : gencond ] ) form the basis of our calculations of sec .
[ sec : results ] .
( [ eq : gczerozerot ] ) relates exactly the zero - bias conductance at @xmath80 to the four spectra @xmath85 at the fermi level , @xmath86 .
these in turn can be obtained exactly in terms of the so - called ` excess charges ' of the dot in the four distinct conduction channels , as now sketched .
the green function @xmath87 is diagonal in spin and orbital indices , and given by @xmath88^{-1}\ ] ] where @xmath89 $ ] is the @xmath90-dependent hybridization function , @xmath91 the effective one - electron energy under @xmath36 ( eqn .
( [ eq : fullh ] ) ) , and @xmath92 is the dot self - energy .
luttinger s integral theorem , @xcite @xmath93 applies separately within _ each _ conduction channel @xmath94 , allowing one to follow the steps of _
e.g. _ [ ] to derive the friedel sum rule : @xcite @xmath95 this relates @xmath96 , the ( fermi level ) phase shift of the conduction electrons in the @xmath94 channel , to the corresponding excess charge given by @xmath97\right\}\ ] ] where @xmath98 [ @xmath99 is the @xmath94 conduction electron spectrum for wavevector @xmath54 in the presence [ absence ] of the dot .
the fermi level value of the spectrum at @xmath80 is readily shown to satisfy @xcite @xmath100 and hence from eqn .
( [ eq : gczerozerot ] ) we obtain @xmath101 the @xmath80 zero - bias conductance is thus related to the excess charges in the four conduction channels ( themselves readily obtained via a thermodynamic nrg calculation ) . in the experimentally relevant limit where @xmath50 is the largest energy scale ,
these excess charges are moreover confined to the dot itself .
one can then approximate @xmath102 by @xmath103 , thereby producing a simple relationship between the dot occupancy and its transport properties . in the @xmath0-symmetric limit ( @xmath104 )
considered previously,@xcite the excess charges are equivalent in all four channels and hence @xmath105 with @xmath106 .
( [ eq : finalfsr ] ) then reduces to@xcite @xmath107 we analyse the model eqn .
( [ eq : fullh ] ) using the numerical renormalization group ( nrg ) .
this technique @xcite has long provided access to numerically - exact results for thermodynamics and , with the recent identification of its complete fock space,@xcite an equally systematic and reliable route to dynamical properties .
the basic approach is detailed in _ e.g. _ [ ] . a logarithmic discretization of the conduction band states is first used to map the hamiltonian onto a countably infinite one - dimensional chain .
the linear chain is diagonalised iteratively , starting from a single site and adding the others one by one .
the key advantage of logarithmic discretization is that the coupling constants along the chain decrease rapidly , and the high - energy states of one iteration can be discarded without affecting the low - energy states retained in later iterations . as such , a fixed number of states can in practice be kept at each iteration , rendering the iterative diagonalization of the hamiltonian numerically tractable .
the information obtained from each iteration allows one to build up the thermodynamics and dynamics of the model .
eigenstates of a given iteration are used to calculate thermodynamics at an appropriately - chosen temperature ( sufficiently low that discarded states of earlier iterations are unimportant , yet sufficiently high that the energy splittings of later iterations are thermally smeared out ) .
this effective temperature decreases exponentially with the iteration number , and hence allows access to thermodynamics on all physical energy scales after only a modest number of iterations .
dynamics of the model are calculated by means of the recent observation @xcite that the set of all _ discarded _ states forms a complete basis of the discretized nrg hamiltonian . by expanding the full density matrix in this basis ,
accurate results for dynamical correlation functions may be obtained over a wide range of frequency and temperature scales.@xcite dynamics of the discrete nrg hamiltonian necessarily arise as a series of isolated poles : in order to capture the behavior of the original continuous model , one then convolves the discrete spectra with an appropriate broadening function on a logarithmic scale ( see _ e.g. _ [ ] ) .
potential artefacts of the discretization process are minimised in three standard ways : dot - lead couplings are premultiplied by the standard @xmath108 factor,@xcite the oliveira ` @xmath10-averaging'@xcite is used to average discrete spectra with different logarithmic discretizations , and the green function is obtained not directly but from its self - energy @xcite , calculated as a ratio of two correlation functions where any remaining discretization effects largely cancel .
the calculations in this work have been obtained with an nrg discretization parameter @xmath109 , exploiting the full @xmath110 symmetry of the model .
we have typically averaged results for five different @xmath10s and have kept the lowest @xmath111@xmath112 states at each iteration .
here we show that a physical understanding of the interplay between kondo physics , so and zeeman couplings , follows readily by considering the isolated dot ( the atomic limit , @xmath113).the latter has been considered in [ ] , from which we take results as required ; denoting the dot states in an obvious notation as @xmath114 , @xmath115 , @xmath116 , and so on .
the energies of the 16 possible isolated dot states follow directly from eqn .
( [ eq : fullh ] ) with @xmath117 , and may be classified by their total occupation number @xmath118 . in the absence of so and zeeman couplings , @xmath119 , the @xmath120 $ ] @xmath121-electron states are degenerate , with energies @xmath122 . on switching on @xmath32 and @xmath22 , the @xmath123 and @xmath124 states
are trivially unaffected , while the @xmath125 electron states are split as shown schematically in fig .
[ fig : alenergies ] .
splittings of the @xmath126 atomic limit states , for finite @xmath22 and @xmath127 ( relative to their values for @xmath128 ) . ] for @xmath129 , so coupling splits both the @xmath130 and @xmath131 states into two degenerate pairs separated by an energy @xmath32 .
the @xmath132 states by contrast are split into three groups , of degeneracies @xmath133 , @xmath134 and @xmath133 , with relative energies @xmath135 , @xmath136 and @xmath32 respectively . switching on the field further splits the states by both spin- and orbital - zeeman effects ( fig .
[ fig : alenergies ] ) , and in the @xmath130 sector a singly - degenerate ground state ( @xmath137 ) arises for all @xmath138 .
the situation is somewhat more complicated in the @xmath132 and @xmath131 sectors , since in both cases competition between so and zeeman effects leads to level crossings in the ground state . in the @xmath132 sector
it is the _ orbital _
zeeman effect that competes with so coupling : the low - field ground state is @xmath139 as favored by the so interaction eqn .
( [ eq : hsoshortb ] ) , while at higher fields the ground state @xmath140 is favored by the orbital zeeman interaction ( eqn .
( [ eq : hb ] ) ) for the experimentally relevant case @xmath141 ; the ground state level crossing occuring at a field @xmath142 in the @xmath131 sector by contrast it is _ spin _ zeeman which now competes with so coupling , producing a level crossing from the low - field ground state @xmath143 to @xmath144 at a field @xmath145 ( @xmath146 ) given by @xmath147 these special values of the field turn out to be central to the kondo physics of the model , as explained below . from the atomic limit energies , the ground state ` phase diagram '
is readily constructed as a function of @xmath22 and @xmath148 ( @xmath149 ) .
[ fig : cbalnoso ] first shows the situation where so coupling is absent , @xmath150 ; solid lines schematic atomic limit ( @xmath151 ) ` phase diagram ' in the absence of so coupling , as a function of @xmath22 and @xmath152 , showing boundaries between ground states of different charge . ] marking the boundaries between states of different ground state charge . for any fixed @xmath22 ,
increasing @xmath148 generates the familiar coulomb blockade ( cb ) staircase .
when @xmath12 lies sufficiently in excess of the fermi level , both orbitals are empty ; and on lowering @xmath12 through the fermi level the total number of electrons on the dot increases stepwise from zero to four .
notice that , for all @xmath22 , fig .
[ fig : cbalnoso ] is symmetric under reflection about the line @xmath153 corresponding to the midpoint of the @xmath132-electron valley ( _ i.e. _ to replacing @xmath154 $ ] ) ; reflecting for @xmath155 the essential equivalence of states with @xmath121 and @xmath156 electrons under a particle - hole transformation ( specifically @xmath157 ) . and at points of degeneracy between @xmath121 and @xmath158 electrons ( fig .
[ fig : cbalnoso ] solid lines ) , there is naturally facile zero - bias sequential tunneling through the dot ( and hence enhanced conductance ) when it is connected to the leads.@xcite degeneracies between states of the _ same _ total charge do not promote sequential tunneling , but are of course vital for the kondo effects arising from coherent cotunneling processes on coupling to the leads ( discussed below ) .
such degeneracies arise at zero - field in the @xmath130 , @xmath132 and @xmath131 valleys , where all states of given @xmath121 are degenerate . for any @xmath159
however , there is an immediate ` transition ' to a singly - degenerate state in each case with maximal @xmath160 .
the above situation changes qualitatively on introduction of so coupling ; as illustrated in fig .
[ fig : cbal ] , as fig .
[ fig : cbalnoso ] , but for the case of a finite so coupling @xmath32 . boundaries between ground states with the same charge are shown by dashed lines . ] showing the generic behavior for @xmath161 and @xmath162 ( albeit for illustration using a somewhat smaller ratio @xmath163 than seen experimentally ) .
this structure is clearly more interesting than fig .
[ fig : cbalnoso ] .
first , the so coupling @xmath32 has a dramatic effect at zero field . in both the @xmath130 and @xmath131 valleys the degeneracy of the ground state at @xmath129
is reduced from 4 to 2 ( as in figs .
( [ fig : alenergies],[fig : cbal ] ) ) ; the degree of freedom associated with this two - fold degeneracy being neither a pure spin nor an orbital pseudospin , but a mixture of the two . in the @xmath132 valley
the effect of @xmath32 is even more severe : the 6-fold degenerate ground state for @xmath150 is replaced by the non - degenerate state @xmath164 , the other five states again lying @xmath165 higher ( _ cf . _ fig .
[ fig : alenergies ] ) . the finite - field ground state level crossings in the @xmath166 electron sectors ( fig .
[ fig : alenergies ] ) , mean as shown in fig .
[ fig : cbal ] ( dashed lines ) that for all @xmath12 when @xmath132 there is a crossing from @xmath139 to @xmath140 at @xmath167@xcite ; and likewise throughout the @xmath131 sector , a crossing from @xmath143 to @xmath144 at @xmath168 . associated with these ground state crossovers
are naturally kinks in the cb steps seen in fig .
[ fig : cbal ] ( solid lines ) , as discussed in [ ] and observed in the sequential tunneling experiments of [ ] on ultraclean cnt dots .
by contrast there is no ground state level crossing in the @xmath169 sector ( fig . [ fig : alenergies ] ) .
hence , as evident in fig .
[ fig : cbal ] , the ` reflection symmetry ' seen in fig .
[ fig : cbalnoso ] for @xmath155 is absent states with @xmath121 and @xmath156 electrons no longer being equivalent under the particle - hole transformation @xmath157 ( arising because @xmath170 under such ) . as mentioned above the significance of degeneracies between same - charge states
is that , on coupling to the leads , their associated low energy degrees of freedom can be screened by many - body kondo effects which enhance conductance through the dot .
@xcite two basic kondo effects may in fact arise,@xcite @xmath0 and @xmath1 .
only when the full model is close to being @xmath0-symmetric does the former arise ( we define ` close to ' shortly ) ; the more common case , occuring for doubly - degenerate atomic limit states where the low - energy effective model maps onto a spin-@xmath171 kondo model , is the @xmath1 kondo effect . one key physical distinction between the two is the low - energy / temperature kondo scale on which they are manifest , generically denoted @xmath172 . in the strongly correlated regime @xmath173 , the kondo scales in the two cases are @xcite [ eq : kondoscales ] @xmath174 modulo prefactors that depend weakly on the bare parameters , such that @xmath175 . with the above in mind , the essential qualitative physics of the model
is readily deduced .
we start with @xmath150 , and consider specifically the zero - temperature limit ( the effect of temperature being simply to smear out the pristine @xmath80 behavior over an energy scale @xmath176 ) .
the appropriate atomic limit picture for @xmath150 is fig .
[ fig : cbalnoso ] . on coupling to the leads the cb steps , and associated zero - bias conductance arising from facile sequential tunneling , are broadened over an energy scale @xmath177.@xcite in the @xmath130 , @xmath178 and @xmath179 electron valleys for @xmath129 , @xmath0 kondo effects
take place@xcite ( effective low - energy @xmath0 kondo , or coqblin - schrieffer , models @xcite being obtained _ via _ a schrieffer - wolff transformation , on retaining the @xmath120 $ ] degenerate dot states in the ground state manifold and the cotunneling processes that connect them ) .
the @xmath0 kondo physics in the @xmath180 valleys will naturally persist at finite fields until @xmath181 ; while for field strengths in excess of the @xmath0 kondo scale , the kondo effect will be destroyed and the conductance will be correspondingly low . and for @xmath155 , the conductance in the @xmath130 and @xmath131 electron valleys as a function of @xmath22 ( or @xmath182 ) will be coincident ; reflecting the equivalence of associated states under the particle - hole transformation discussed above . on introducing a finite @xmath32 ,
the key quantity in determining whether @xmath0 or @xmath1 kondo physics prevails is the ratio @xmath183 ( with @xmath184 the @xmath0 scale in the absence of so coupling ) .
for @xmath185 the zero - field @xmath0 kondo effects described above are still favorable , since the kondo stabilization energy outweighs the splittings of the atomic limit states in fig .
[ fig : alenergies ] . for these small @xmath32s ,
one thus expects the physics to be essentially unchanged from the @xmath150 limit .
only when @xmath32 becomes comparable to @xmath184 will it have a noticeable effect .
for @xmath186 by contrast the appropriate starting picture is now fig .
[ fig : cbal ] , and @xmath0 kondo effects no longer arise .
the @xmath132 valley will not exhibit any kondo effect at zero - field , since its ground state is singly degenerate .
the @xmath130 and @xmath131 valleys for @xmath129 will however display @xmath1 kondo effects involving their doubly - degenerate ground states ( fig .
[ fig : cbal ] ) : an effective low - energy kondo model of @xmath1 form obviously arises in each case , under a schrieffer - wolff transformation retaining the appropriate degenerate pair of dot states indicated in fig .
[ fig : cbal ] . for @xmath138 ,
the zero - field kondo effects in the @xmath187 valleys will again ` spill over ' into the @xmath22-plane by an amount of order their @xmath1 kondo scales ( rather less than in the @xmath0 case , as above ) . in addition however , the level crossings occurring at finite fields in both the @xmath132 @xcite and @xmath131 valleys ( fig .
[ fig : cbal ] ) means that additional @xmath1 kondo effects will now occur at fields @xmath167 and @xmath188 respectively ( the two - fold ground state degeneracy of the free dot states at either field generating an @xmath1 kondo model under schrieffer - wolff ) .
these will be discernable as long as they are not subsumed by the zero - field @xmath1 kondo effect .
this is clearly not an issue in the @xmath132 valley , no zero - field kondo effect occurring here anyway for @xmath186 ; while in the @xmath131 valley it requires @xmath189 ) @xmath190 , readily seen to be satisfied since @xmath191 . in the @xmath130 valley by contrast , the absence of a ground state level crossing at finite field ( fig .
[ fig : cbal ] ) means that the @xmath1 kondo effect arising here at zero - field will simply be steadily destroyed with increasing @xmath22 , dying out on a scale of order @xmath192 . and
since the @xmath130 and @xmath131 electron valleys in particular exhibit distinct behavior as a function of field , then , as for the atomic limit states themselves , the @xmath193 symmetry of the conductance as a function of @xmath194 is again absent for @xmath195 .
the above considerations are purely qualitative , and we have analysed the model in detail _ via _ nrg , over a large parameter space . here we present a selection of results , focusing in particular on parameter regimes applicable to experiment . specifically , we make comparison to two experimental works : makarovski _ et al . _
@xcite ( denoted ` m ' ) and jarillo - herrero _ et al . _
@xcite ( ` jh ' ) . the former device is somewhat more strongly correlated than the latter ( although in both cases the ratio @xmath196 is sufficiently large to generate non - trivial kondo behavior ) , so we can compare theory to experiment in two distinct physical regimes .
details of how the model parameters are chosen will be given at appropriate points in the following .
it will also be convenient to define and use the reduced parameters : @xmath197 , @xmath198 , @xmath199 and @xmath200 .
[ fig : zbcondv ] shows the @xmath129 zero - bias conductance as a function of the dimensionless gate voltage @xmath201 ( sec .
[ sec : model ] ) , for a range of temperatures @xmath202 and three different so coupling strengths .
here we take an interaction @xmath203 , so that the relative widths of the cb peaks for @xmath204 are in line with the experiments of m ( _ cf . _ the discussion in [ ] ) .
( color online . ) zero - bias conductance @xmath205 as a function of dimensionless gate voltage @xmath42 , at @xmath129 and for @xmath206 .
each panel shows the conductance for the range of temperatures indicated , with so coupling strengths ( a ) @xmath207 , ( b ) @xmath208 and ( c ) @xmath209 . ]
[ fig : zbcondv](a ) shows the @xmath150 case , considered in [ ] .
the @xmath80 conductance ( dotted line ) evolves stepwise with @xmath42 , and follows eqn .
( [ eq : fsrsu4 ] ) as a function of @xmath210 ; with a stepwise increase in @xmath210 itself as the dimensionless gate voltage is increased.@xcite the latter is a result of the relatively large @xmath196 , leading to strong charge quantization on the dot except when @xmath12 is within @xmath177 of the atomic limit charge - degeneracy points at @xmath211 , @xmath212 , @xmath213 . on increasing @xmath182 the zero - bias conductance is rapidly eroded towards the centers of the @xmath214 , @xmath178 and @xmath179 electron valleys , resulting in the familiar coulomb - blockade valley structure with conductance peaks around the atomic limit charge - degeneracy points .
the temperature scale over which the erosion takes place is of course @xmath184 , given by eqn .
( [ eq : kondoscales ] ) and obtained numerically @xcite as @xmath215 [ @xmath216 for @xmath217 [ @xmath133 ] . for @xmath219 , the hwhm of the coulomb blockade peaks are of order @xmath177 ( not _ precisely _
@xmath69 due@xcite to electron interactions ) ; while for @xmath220 the cb peaks simply broaden to become of width @xmath176 instead .
[ fig : zbcondv](b ) shows the effect of a non - zero so coupling of order @xmath221 , such that so coupling competes with the @xmath0 kondo effects in the centers of the coulomb blockade valleys .
the @xmath80 conductance is slightly eroded around @xmath222 , but is qualitatively unchanged elsewhere . upon increasing @xmath182 ,
it is clear that once @xmath223 , the conductance appears essentially identical to the @xmath150 case fig .
[ fig : zbcondv](a ) , as one expects physically .
on now considering @xmath186 ( but still small relative to the non - universal scale @xmath69 ) , the situation changes to that of fig .
[ fig : zbcondv](c ) . here
the kondo effect in the two - electron coulomb - blockade valley at @xmath222 is destroyed at @xmath80 as expected ( sec .
[ sec : atomic ] ) , and the conductance remains rather low for all temperatures shown .
for the @xmath224 and @xmath179 electron cb valleys by contrast , the @xmath80 conductance is still @xmath225 ( as for the @xmath0 symmetric limit @xmath155 , which follows in that case from eqn .
( [ eq : fsrsu4 ] ) with @xmath226 @xmath227 for the center of the @xmath224 @xmath227 electron cb valley ) .
this behavior for _ large _
@xmath228 is now however symptomatic of the @xmath1 kondo effect arising in that case ; as readily understood using the general result eqn .
( [ eq : finalfsr ] ) for @xmath229 , considering the @xmath224 case explicitly .
recall ( fig .
[ fig : cbal ] ) that the free dot ground state for @xmath224 is the degenerate pair @xmath137 and @xmath230 ( which generate @xmath1 kondo under schrieffer - wolff on cotunneling to the leads , sec .
[ sec : atomic ] ) , for which @xmath231 and @xmath232 ; and since @xmath233 as noted in sec .
[ sec : fsr ] , eqn .
( [ eq : finalfsr ] ) gives directly @xmath234 .
note however that although the @xmath80 conductance for @xmath224 or @xmath179 barely discriminates between @xmath0 ( @xmath155 , fig .
[ fig : zbcondv](a ) ) and @xmath1 ( @xmath235 , fig .
[ fig : zbcondv](c ) ) , the erosion of conductance with temperature occurs more rapidly in the latter case occurring naturally on the @xmath1 kondo scale @xmath236 ( with @xmath237 in fig .
[ fig : zbcondv](c ) ) .
the full @xmath182-dependence of the conductance in the centers of the cb valleys is shown in fig .
[ fig : zbcondt ] , @xmath229 _ vs _ @xmath238 ( on a log - scale ) for a range of @xmath239 values , and for @xmath240 ( fig .
[ fig : zbcondt](a ) ) and @xmath224 ( fig .
[ fig : zbcondt](b ) , @xmath224 and @xmath179 being equivalent by symmetry for @xmath129 , see _
e.g. _ fig .
[ fig : alenergies ] ) .
( color online . )
zero bias conductance @xmath205 _ vs _
@xmath238 for various so couplings @xmath239 as indicated , at @xmath129 with @xmath206 . for ( a ) the center of the @xmath240-electron cb valley ; ( b ) the center of the @xmath224 valley . for @xmath129 ,
the behavior for @xmath241 is identical to ( b ) by symmetry ( see fig . [
fig : alenergies ] ) . ] results for @xmath242 have been considered in [ ] : in each valley the strong conductance enhancement due to coherent @xmath0 kondo transport is evident in the ` kondo plateau ' for temperatures @xmath243 ( although note that the universal scaling forms of @xmath244@xmath245 in the two valleys differ quantitatively,@xcite reflecting the two distinct @xmath0 kondo effects that arise therein ) . on a temperature scale @xmath246 ,
a conductance shoulder is also evident , corresponding to incoherent sequential tunneling transport . in the @xmath240-electron valley ( fig .
[ fig : zbcondt](a ) ) the kondo plateau is progressively destroyed on increasing the so coupling @xmath239 , as all but the lowest ( non - degenerate ) atomic limit states become projected out of the low - energy manifold ( see fig .
[ fig : alenergies ] ) . for @xmath32
sufficiently large compared to @xmath247 , a peak is seen to emerge in the conductance at @xmath248 ( in practice @xmath249 ) naturally so , this being the energy gap to higher so - split states ( fig .
[ fig : alenergies ] ) , and mixing in of which enhances the conductance . and in all cases shown , the high - temperature ( @xmath250 ) behavior is entirely coincident regardless of @xmath32 . as expected from the discussion above , the @xmath182-dependence of the conductance upon increasing the so coupling @xmath239 in the center of the one - electron valley ( fig .
[ fig : zbcondt](b ) ) , shows clearly a crossover from the @xmath0 behavior arising for @xmath251 , to the @xmath1 behavior arising asymptotically for @xmath235 .
this limiting form occurs in practice for @xmath252 ( _ i.e. _ @xmath253 ) , such that a further increase in @xmath239 naturally leaves the @xmath182-dependence of @xmath229 unchanged , as seen in the figure .
we turn now to finite field , fixing the so coupling @xmath254 and considering @xmath255 as a function of the dimensionless gate voltage @xmath256 and field @xmath257 .
we first consider @xmath80 , in terms of which the finite-@xmath182 behavior is readily understood .
the @xmath207 conductance as a function of field@xcite is shown in fig .
[ fig : condmaps](a ) .
( color online . ) @xmath80 zero - bias conductance maps , @xmath205 as a function of field @xmath200 and dimensionless gate voltage @xmath42 , for ( a ) @xmath242 and ( b ) @xmath258 .
bare parameters here are @xmath259 and @xmath260 . ] at @xmath129 it has the stepwise form seen in fig .
[ fig : zbcondv ] , with a maximal conductance plateau of @xmath261 in the center of the two - electron valley ( @xmath217 ) and plateaux of @xmath262 in the centers of the one- and three - electron valleys . for @xmath263 ,
the associated @xmath0 kondo effects are progressively destroyed : four distinct cb peaks instead emerge , centered along the lines of atomic limit charge degeneracy ( _ cf . _ fig .
[ fig : cbalnoso ] ) .
the @xmath0 kondo behavior at @xmath129 persists over a finite @xmath22-range , but is eventually destroyed for @xmath264 .
on introducing a finite so coupling @xmath265 , the picture changes to that of fig .
[ fig : condmaps](b ) ( _ cf .
_ fig .
[ fig : cbal ] ) .
the coulomb blockade lines now show the expected kinks at the fields @xmath266 and @xmath267 , while for sufficiently high @xmath268 , so coupling is of course negligible and the behavior approaches that of fig .
[ fig : condmaps](a ) .
( color online . )
slices through the @xmath224 , @xmath240 and @xmath269 valleys of fig .
[ fig : condmaps](b ) , showing the evolution of the conductance @xmath270 with temperature @xmath238 as indicated . ] of primary interest here is the effect of so coupling on the kondo physics .
as seen earlier , at @xmath129 the so coupling destroys @xmath0 kondo in the two - electron valley , eliminating the conductance maximum in this region .
it also reduces the one- and three - electron zero - field @xmath0 kondo effects to @xmath1 ( the latter apparent in fig .
[ fig : condmaps](b ) from a clear reduction in the field strength required to destroy the kondo peak ) .
although so coupling thus has a destructive influence on the kondo effects for @xmath129 , it leads as discussed in sec .
[ sec : atomic ] to two finite - field @xmath1 kondo effects ( fig .
[ fig : condmaps](b ) ) when level crossings occur in the atomic limit ground states .
both lead to a significant enhancement of the @xmath80 conductance , of order @xmath271 and extending over field ranges @xmath272 , resulting in marked differences between fig .
[ fig : condmaps](a ) and fig .
[ fig : condmaps](b ) .
the effect of temperature on conductance maps is best seen by taking slices through the centers of the cb valleys in the @xmath273 plane .
the center of the two - electron valley is ( by symmetry ) @xmath274 for all @xmath22 , while for the one- and three - electron valleys we take the trajectories @xmath275 and @xmath276 respectively ( which approach the centers of the cb valleys in the large-@xmath22 limit ) .
the results are shown in fig .
[ fig : cbvalleyt ] . in the two - electron valley
the only kondo effect is the @xmath1 kondo ` revived ' at the finite field @xmath167 ( @xmath277 here).@xcite the corresponding @xmath278 , and hence on increasing @xmath182 the kondo effect is in essence destroyed by @xmath279 .
directly analogous comments apply to the zero - field kondo effects arising in the one- and three - electron valleys ; and , for the latter case , to the additional finite - field @xmath1 kondo effect revived at @xmath188 ( @xmath280 here ) .
we also note that the clear so - induced asymmetry between the one- and three - electron valleys persists even for temperatures @xmath281 , where the finite - field kondo peak at @xmath282 is itself thermally washed out : relatively small though it is , the conductance in the @xmath269 valley appreciably exceeds that in the @xmath224 valley over a wide @xmath283-interval .
thus far we have focussed on the strongly correlated regime where @xmath284 .
this leads to a pristine separation of energy scales : the cb peaks are separated by many times their widths , and the kondo scales are exponentially smaller than @xmath69 . on moving to a more moderately correlated regime , the energy scales naturally begin to merge , but the essential situation remains the same .
an example is shown in fig . [
fig : condmapwk ] , where now @xmath285 and @xmath286 ( and @xmath39 has been reduced slightly to @xmath287 ) .
( color online . ) as fig . [
fig : condmaps](b ) , but for a reduced coupling strength @xmath288 , with @xmath289 and @xmath290 . ] for @xmath285 , the resultant @xcite zero - field @xmath1 kondo scale for example is @xmath291 , and hence @xmath292 , @xmath32 and @xmath69 all comparable . as seen from fig .
[ fig : condmapwk ] , in comparison to the more strongly correlated fig .
[ fig : condmaps ] this generates a more pronounced asymmetry between the one- and three - electron valleys on increasing @xmath22 , in particular with the cb peaks in the three - electron valley brought closer together .
the results shown in fig . [ fig : condmapwk ] agree well with the experimental results of jh @xcite for the four - electron shell centered on @xmath293 , @xcite as evident from direct comparison with fig .
2(b ) of jh in the interval @xmath294 . and
more significantly , they provide a natural explanation for the observations , as arising from the interplay between spin - orbit and kondo physics .
the bare parameters employed in fig . [ fig : condmapwk ] are themselves consistent with the jh experiment . from fig .
2(b ) of jh ( the @xmath295-@xmath296 line therein ) one readily identifies the experimental @xmath297 , and likewise the experimental ratio @xmath298 ( from the @xmath295-@xmath296 and @xmath299-@xmath300 lines ) . since @xmath301 ( eqn .
( [ eq : bsdef ] ) ) , and @xmath302 ( we take @xmath303 ) , the experimental @xmath145 gives the so coupling constant as @xmath304 which we note is in line with that measured recently in the cnt experiments of [ ] via sequential tunneling spectroscopy at finite bias .
fitting the atomic limit cb peaks to the jh experimental data gives @xmath305 and hence @xmath306 , in good agreement with the heights of the cb diamonds in fig .
2 of jh ; while the ratio @xmath196 is estimated straightforwardly by comparing the widths of the cb peaks to their separation . finally , the experimental temperature @xmath307 is sufficiently small compared to the other scales in the problem , that one can set @xmath80 with impunity in the nrg calculations . as above
, we consider the behavior seen in fig .
[ fig : condmapwk ] to be in striking agreement with fig .
2(b ) of jh in the region @xmath294 .
we also point out that the experiment deviates from our calculation above a gate voltage @xmath308 .
this arises simply because the levels of the _ adjacent _ four - electron shell in experiment are brought into play by the magnetic field , and at sufficiently high @xmath22 ` interfere ' with those arising from the shell considered .
this is naturally not taken into account in the model ( although it would be straightforward to incorporate ) .
so far we have considered the zero - bias conductance as a function of gate voltage , magnetic field and temperature .
experimentally there is another ` knob to turn ' , the source - drain bias voltage @xmath82 . as explained in sec .
[ sec : theory ] , this we handle approximately using eqn .
( [ eq : gencond ] ) , which relates the finite - bias conductance to the frequency dependence of the equilibrium single - particle spectra @xmath81 .
our main interest here is how finite fields and so coupling affect the low - energy kondo behavior of the conductance .
when @xmath309 , all four @xmath81s share a common @xcite @xmath0 kondo resonance in each of the @xmath214 , @xmath178 and @xmath179 cb valleys . on introducing a finite @xmath22 and @xmath32 , and thus lowering the symmetry to @xmath110 ,
each @xmath81 instead possesses a distinct kondo resonance . on application of a field , these resonances shift away from the fermi level @xmath310 by different amounts , and at high fields in particular the four resonances are sufficiently well separated that the combined spectrum @xmath311 contains four separate peaks.@xcite we also emphasise at this point that , despite occasional naive belief to the contrary , the field - induced shifts of the kondo resonance are _ not _ simple ` zeeman splittings ' : they have in fact a non - linear field dependence brought about by the strong interactions on the dot , which can either underestimate or overestimate the zeeman splitting depending on the strength of the field ( see _ e.g. _ [ ] and refs therein for a discussion of the @xmath1 anderson model ) . bearing the above in mind ,
we consider ( fig .
[ fig : cbf ] ) finite - bias differential conductance maps , @xmath312 as a function of @xmath57 and gate voltage @xmath42 , with each taken at fixed field .
again we start at @xmath80 , moving to finite-@xmath182 later when comparing to the experiments of m @xcite . taking the limit @xmath313 in eqn .
( [ eq : gencond ] ) shows that @xmath314 is proportional to the average of @xmath315 , _
i.e. _ the ( approximate ) conductance amounts to a symmetrized combination of the total single particle spectrum of the dot .
the @xmath128 behavior has been described in [ ] ( fig . 5 therein ) .
two distinct features arise : the narrow zero - bias @xmath0 kondo ridges produced by coherent many - body tunneling , and finite bias coulomb blockade diamonds generated by incoherent sequential tunneling .
the former occur only below @xmath182s of order @xmath184 and are likewise destroyed by the source - drain bias when @xmath316 becomes of the same order , while the latter are of width @xmath317 and hence rather more robust .
figure [ fig : cbf ] shows the behavior for a finite @xmath318 at @xmath80 , for a range of field strengths .
the @xmath129 conductance is shown in fig .
[ fig : cbf](a ) .
the coulomb blockade diamonds are essentially unchanged from the @xmath319 limit@xcite since @xmath320 , and the form of the zero - bias conductance is as discussed in relation to fig .
[ fig : zbcondv ] : on switching on @xmath32 , the one- and three - electron valleys ( @xmath321 and @xmath179 ) show @xmath1 kondo effects instead of @xmath0 , while the conductance in the two - electron valley at @xmath322 is substantially reduced .
we now see from fig .
[ fig : cbf](a ) that the reduction of the zero - bias conductance in the two - electron valley in fact reflects a splitting of the kondo resonance in the symmetrized spectrum : two narrow conductance peaks are seen to arise for @xmath322 when @xmath323 . on slightly increasing the field to @xmath167 , fig .
[ fig : cbf](b ) , the behavior around @xmath324 changes .
first , the @xmath1 kondo resonances in the one- and three - electron valleys split ( in analogy to the well known behavior of the @xmath1 aim in a magnetic field ) . in the two - electron valley at @xmath167
by contrast , the orbital @xmath1 kondo effect described earlier arises , and as such the splitting of the kondo resonance here is reduced to zero .
the next ` special ' value of the field is @xmath188 , shown in fig .
[ fig : cbf](c ) . here
the @xmath1 spin kondo effect takes place in the three - electron valley , whence the splitting of the kondo resonance seen in fig .
[ fig : cbf](b ) is reduced to zero .
in addition , two faint ` shoulders ' at a small finite bias can just be made out .
these are the beginnings of the separation of the total spectrum @xmath325 into four separate components at high field ( as mentioned above ) , which we discuss in more detail below .
finally , for @xmath326 the atomic limit ground states in all valleys are unchanged with increasing @xmath22 , and the low - energy peak splittings in the @xmath82-dependence of the conductance all increase monotonically .
moving from ( d ) to ( e ) and ( f ) in fig . [
fig : cbf ] , we see that once the field becomes of order @xmath69 the kondo peaks simply merge with the cb diamonds ( or the hubbard satellites in single - particle spectra terminology ) , and the latter themselves begin to split from then on . at this point
we make our first comparison with the experiment of m @xcite . keeping the ratio @xmath327 , we take @xmath328 such that the resulting @xmath329 is in good agreement with the heights of the cb diamonds in fig . 2 of m.
our choice of @xmath330 then corresponds to @xmath331 , which is of the same order of magnitude as that measured in another device.@xcite and we now consider @xmath332 and @xmath333 , to be in line with the experimental temperature and field , @xmath334 and @xmath335 . using a colormap similar to m , and taking @xmath336 , we obtain fig . [
fig : cbft ] , as fig .
[ fig : cbf ] but with @xmath337 , @xmath338 , and a colormap chosen to be similar to fig . 2 of [ ] .
the arrows indicate the positions of low - energy peaks , see text .
taking @xmath339 , the @xmath316 axis extends from @xmath340 to @xmath341 mev , @xmath342 and @xmath343 , in agreement with fig . 2 of [ ] .
] which is to be compared with fig .
2 of m. our results are in good agreement with the experiment . in particular
, we note that the low - energy features identified in the experimental paper@xcite are reproduced by the calculations . in the two - electron valley
we see a circular region of reduced conductance , while in the one- and three - electron valleys we capture two and one low - energy peak , respectively ( marked by arrows in fig .
[ fig : cbft ] ) . on comparing fig .
[ fig : cbft ] to the @xmath80 results in fig .
[ fig : cbf ] , the apparent single peak in the center of the three - electron valley is in fact seen to be a consequence of thermal broadening : at zero temperature in the three - electron valley , a single peak centered at zero - bias occurs only at the special fields of @xmath129 and @xmath188 as explained above ( at any other field , this zero - bias peak is always split ) .
while a similar splitting arises also in the one - electron valley , it increases monotonically from @xmath129 and is hence somewhat larger than that of the three - electron valley . as a result , the two peaks in the one - electron valley remain separate at the particular temperature @xmath338 ( @xmath344 ) used in fig . [
fig : cbft ] , while the two peaks in the three - electron valley are merged into one .
if the experiment had been performed at a sufficiently lower temperature , we would expect two low - energy peaks in both the one- and three - electron valleys . to compare further with experiment , we now fix the gate voltage to lie at the centers of the one- , two- and three - electron valleys ( _ cf . _ fig .
[ fig : cbvalleyt ] ) and consider the conductance as a function of source - drain bias and field .
the results are shown in fig .
[ fig : valleyvb](a)(c ) as colormaps , and in fig .
[ fig : valleyvb](d)(f ) as slices at fixed @xmath345 , @xmath3 , @xmath346 , @xmath347 , @xmath348 . for clarity ,
the data in fig .
[ fig : valleyvb](d)(f ) have been shifted vertically by @xmath349 to separate the individual lines .
the figure is to be compared with fig . 3 of m. we see in fig .
[ fig : valleyvb](a)(c ) the evolution of the low - energy conductance peaks with increasing field , which again agree rather well with experiment . in the one - electron valley ( fig .
[ fig : valleyvb](a ) ) the single kondo peak at @xmath129 is seen to split into the four spectral features highlighted in m ; while , as discussed earlier , in the three - electron valley ( fig .
[ fig : valleyvb](c ) ) for sufficiently - small @xmath22 the two lowest - energy peaks are merged by thermal broadening . for larger @xmath22 ( @xmath350 here ) the two peaks in the three - electron valley
do eventually separate in our calculations , as expected on physical grounds .
this splitting is difficult to see in the experiment due to the neighboring four - electron shell being brought into play at high field , but should be observable in a similar device with a larger energy separation between shells .
the destruction of the kondo effect in the two - electron valley with increasing @xmath22 , as observed in m , is also clearly seen ( fig . [
fig : valleyvb](b ) ) , first as a splitting of the @xmath129 kondo resonance which then rapidly enlarges to leave an almost rectangular - shaped ` hole ' in the conductance around @xmath351 ( fig .
[ fig : valleyvb](e ) ) . note also that the data slices shown in fig .
[ fig : valleyvb](d f ) are symmetrical about @xmath352 , reflecting the assumption in the calculations of a perfectly symmetrical voltage split between the leads ( sec .
[ sec : theory ] ) .
their experimental counterparts in fig
. 3 of m for the one- and two - electron valleys ( which are not appreciably affected by ` overlap ' with a higher shell ) are somewhat asymmetrically disposed about @xmath352 .
this can in fact be reproduced in calculation by parametrising a small degree of asymmetry into the voltage split ( without affecting the essential quality of fig .
[ fig : valleyvb](a c ) ) , although we do not pursue it further here .
in this paper we have studied the effect of so coupling in carbon nanotube quantum dots , by applying the nrg to a modified @xmath0 anderson impurity model ( aim ) .
our main focus has been the case in which the so coupling is comparable to or exceeds the @xmath0 kondo scale , since here the two effects interplay and compete , leading to a rich range of physical behavior .
the differential conductance over a wide parameter space has been calculated as a function of gate voltage , magnetic field and temperature , in order to elucidate the key physics of the model .
we have moreover shown that the inclusion of so coupling accounts for a number of important experimental observations in the works of jarillo - herrero _ et al . _
@xcite and makarovski _ et al . _
@xcite , the origin of which stems directly from the interplay between so and kondo physics . to conclude , we comment on the suitability of the ` pure ' @xmath0 aim as a model for carbon nanotube quantum dots . in [ ] , experimental data of makarowski _ et al .
_ at zero - field@xcite was found to be in good agreement with the @xmath0 anderson model _ without _ including so coupling , over a wide range of @xmath196 . given the results of the present paper , one naturally asks : why ?
let us first summarise the experiment .
the conductance of the experimental device@xcite was measured , at several fixed temperatures , as a function of the applied gate voltage .
the latter was swept over a sufficiently wide range that four different electron shells ( ` groups i iv ' ) were brought through the fermi level , one at a time .
a consequence of varying the gate voltage by this relatively large amount was that the tunnel couplings to the dot ( @xmath69 ) varied from one shell to the next .
as a result , the group i data were described by an @xmath0 model with @xmath353 ( and @xmath354 ) , groups ii and iii were more consistent with @xmath355 ( @xmath356 ) , while group iv had @xmath357 ( @xmath358 ) . to understand why these data could be described by the @xmath0 model , we note that ( a ) the experimental ` base ' temperature was @xmath359 and ( b ) the kondo scales for groups ii
iv are all in excess of @xmath360 ( see fig .
3(b ) of [ ] ) .
assuming the so coupling to be comparable to the value @xmath361 considered above , groups ii
iv _ can _ then be described by a pure @xmath0 anderson model ( for essentially all temperatures ) , since each has a @xmath184 appreciably in excess of the so coupling .
while the latter is not the case for the group i shell @xcite of [ ] , data were only obtained down to a temperature of order @xmath32 where , according to sec .
[ sec : zbb0 ] , the effects of so coupling can not be seen in the zero - field behavior alone . only by examining the behavior of the group i shell in a magnetic field ( as considered here ) , or by measuring its zero - field conductance down to a rather lower temperature on the order of @xmath362 or so , can the effects of so coupling be observed . | we investigate the influence of spin - orbit coupling on the kondo effects in carbon nanotube quantum dots , using the numerical renormalization group technique .
a sufficiently large spin - orbit coupling is shown to destroy the @xmath0 kondo effects at zero magnetic field , leaving only two @xmath1 kondo effects in the one- and three - electron coulomb blockade valleys .
on applying a finite magnetic field , two additional , spin - orbit induced @xmath1 kondo effects arise in the three- and two - electron valleys . using physically realistic model parameters
, we calculate the differential conductance over a range of gate voltages , temperatures and fields .
the results agree well with measurements from two different experimental devices in the literature , and explain a number of observations that are not described within the standard framework of the @xmath0 anderson impurity model . |
the ytterbium atom has two fermionic and five bosonic isotopes , a @xmath9 ground state , a long - lived metastable @xmath10 state , and transitions at convenient wavelengths for laser cooling and trapping .
all this makes yb a superb candidate for a variety of applications such as development of optical atomic clocks @xcite , study of degenerate quantum gases @xcite , quantum information processing @xcite , and studies of fundamental symmetries @xcite .
the best limit to date on the value of the electron electric - dipole moment ( edm ) which constrains extensions of the standard model of electroweak interactions , was obtained using the ybf molecule @xcite .
ybrb and ybcs molecules have also been proposed for searches for the electron edm @xcite since they can be cooled to very low temperatures and trapped in optical dipole traps , leading to very long coherence times in comparison to molecular beam edm experiments .
yb is of particular interest for studying quantum gas mixtures @xcite .
significant progress has been achieved in studying the properties of yb - yb photoassociation spectra at ultralow temperatures @xcite .
photoassociation spectroscopy has been performed on bosons @xcite and fermions @xcite .
the use of optical feshbach resonances for control of entangling interactions between nuclear spins of @xmath11yb atoms for quantum information processing applications has been proposed in @xcite .
a p - wave optical feshbach resonance using purely long - range molecular states of a fermionic isotope of ytterbium @xmath11yb was demonstrated in @xcite .
recent work @xcite theorizes that the case of @xmath12yb may have sufficiently small direct background interaction between the atoms to support two bound states that represent attractively and repulsively bound dimers occurring simultaneously .
the excited molecular states asymptotically connected to the @xmath13 separated yb atom limit were investigated by takasu _ et .
_ in @xcite .
they reported the successful production of a subradiant @xmath14 state of a two - atom yb system in a three - dimensional optical lattice .
the properties of the long - range potential were studied and the van der waals coefficients @xmath15 , @xmath3 , and @xmath6 were predicted .
however , fit of the @xmath3 and @xmath6 coefficients for the @xmath14 state was rather uncertain , with strong correlation between the @xmath3 and @xmath6 fit parameters @xcite .
knowledge of the @xmath3 and @xmath6 long - range interaction coefficients in yb - yb dimers is critical to understanding the physics of dilute gas mixtures .
recently , we evaluated the @xmath3 coefficient for the yb - yb @xmath16 dimer and found it to be @xmath17 @xcite , in excellent agreement with the experimental result @xmath18 @xcite .
however , the same method can not be directly applied to the calculation of the van der waals coefficients with yb - yb @xmath19 dimer owing to the presence of the @xmath20 decay channel . in this work ,
we develop the methodology for an accurate evaluation of the van der waals coefficients of dimers involving excited state atoms with a strong decay channel to the ground state and evaluate @xmath3 and @xmath6 coefficients of particular experimental interest .
we carefully study the uncertainties of all quantities calculated in this work so the present values can be reliably used to analyse existing measurements and to facilitate planning of the future experimental studies .
the methodology developed in this work can be used for evaluation of van der waals coefficients in a variety of systems .
we investigate the molecular potentials asymptotically connecting to the @xmath21 atomic states .
the wave function of such a system constructed from these states is @xmath22 where the index i(ii ) describes the wave function located on the center i(ii ) and @xmath23 . here , the @xmath24 is the projection of the appropriate total atomic angular momentum @xmath25 on the internuclear axis .
we assume that @xmath26 is a good quantum number for all calculations in this work ( hund s case ( c ) ) .
the molecular wave functions can be obtained by diagonalizing the molecular hamiltonian @xmath27 in the model space . here ,
@xmath28 and @xmath29 represent the hamiltonians of the two noninteracting atoms and @xmath30 is the residual electrostatic potential defined as the full coulomb interaction energy in the dimer excluding interactions of the atomic electrons with their parent nuclei . unless stated otherwise , throughout this paper we use atomic units ( a.u . ) ; the numerical values of the elementary charge , @xmath31 , the reduced planck constant , @xmath32 , and the electron mass , @xmath33 , are set equal to 1 .
the atomic unit for polarizability can be converted to si units via @xmath34 [ hz/(v / m)@xmath35=2.48832@xmath36 ( a.u . )
, where the conversion coefficient is @xmath37 , @xmath38 is the bohr radius and @xmath39 is the dielectric constant .
the potential @xmath40 may be expressed as an expansion in the multipole interactions : @xmath41where @xmath42 are given by @xcite @xmath43here , @xmath44 and the multipole spherical tensors are @xmath45where the summation is over atomic electrons , @xmath46 is the position vector of electron @xmath47 , and @xmath48 are the reduced spherical harmonics @xcite .
we now restrict our consideration to the dipole - dipole and dipole - quadrupole interactions . introducing designations @xmath49 , @xmath50 , @xmath51 , and @xmath52
, we obtain from eq .
( [ vll ] ) : @xmath53 , \nonumber \label{vdd}\end{aligned}\]]where the dipole and quadrupole weights are @xmath54numerically , @xmath55 @xmath56 and @xmath57 .
the energy @xmath58 , where @xmath59 and @xmath60 are the atomic energies of the @xmath61 and @xmath62 states , is obtained from @xmath63 the molecular wave function @xmath64 can be formed as a linear combination of the wave functions given by eq .
( [ yb_wf ] ) .
@xmath64 poses a definite gerade / ungerade symmetry and definite quantum number @xmath26 .
it can be represented by @xmath65 where we set @xmath66 for ungerade symmetry and @xmath67 for gerade symmetry .
we have taken into account that the states @xmath68 and @xmath69 that are of interest to the present work are the opposite parity states of yb atom ( when @xmath70 ) .
applying the formalism of rayleigh - schredinger perturbation theory in the second order @xcite and keeping the terms up to @xmath71 in the expansion of @xmath40 we obtain the dispersion potential in two - atom basis : @xmath72 , \label{ur0}\end{aligned}\]]the intermediate molecular state @xmath73 with unperturbed energy @xmath74 runs over a _
complete _ set of two - atom states , excluding the model - space states , eq .
( [ yb_wf ] ) . the dispersion potential can be approximated as @xmath75 the first - order correction , which is determined by the first term on the right - hand side of eq .
( [ ur0 ] ) , is associated with the @xmath15 coefficient in eq .
( [ ur1 ] ) . for the states considered in this work
, this coefficient is nonzero only for the molecular potential asymptotically connecting to the @xmath13 atomic states .
it depends entirely on the reduced matrix element ( me ) of the electric - dipole operator @xmath76 and is given by a simple formula @xmath77 specifically , @xmath78 where the upper / lower sign corresponds to gerade / ungerade symmetry .
the second - order corrections , associated with the @xmath3 and @xmath6 coefficients , are given by the second and third terms on the r.h.s .
of eq .
( [ ur0 ] ) , @xmath79 where @xmath80 and the complete set of doubled atomic states satisfies the condition @xmath81 after angular reduction , the @xmath3 coefficient can be expressed as @xmath82 where @xmath83 ^ 2,\nonumber \\ & & x_{j_\alpha j_\beta } = \sum_{\alpha,\beta \neq a , b } \frac{|\langle a||d||\alpha \rangle|^2 \ , |\langle b ||d|| \beta \rangle|^2 } { e_\alpha - e_a + e_\beta- e_b}\end{aligned}\ ] ] with fixed @xmath84 and @xmath85
. if @xmath68 and @xmath69 are the spherically symmetric atomic states and there are no downward transitions from either of them , the @xmath3 and @xmath6 coefficients for the @xmath86 dimers are given by well known formulas ( see , e.g. ,
@xcite ) @xmath87 where the coefficients @xmath88 ( @xmath89 ) are quadratures of electric - dipole , @xmath90 , and electric - quadrupole , @xmath91 , dynamic polarizabilities at an imaginary frequency : @xmath92 for the yb
yb @xmath93 dimer considered in this work , the expressions for @xmath3 and @xmath6 are more complicated due to the angular dependence , the @xmath94 decay channel and non - vanishing quadrupole moment of the @xmath95 state .
after some transformations , we arrive at the following expression for the @xmath96 coefficient in the @xmath93 case : @xmath97 where the angular dependence @xmath98 is represented by @xmath99 with the dipole weights @xmath100 given by eq .
( [ wmu ] ) and @xmath101 .
it is worth noting that @xmath98 ( and , consequently , the @xmath3 coefficients ) do not depend on gerade / ungerade symmetry .
the quantities @xmath102 for the @xmath93 dimer are given by @xmath103 where @xmath104 and @xmath105 and @xmath106 is defined below .
the possible values of the total angular momentum @xmath107 are 0 , 1 , and 2 ; @xmath108 is the electric - dipole dynamic polarizability of the @xmath9 state at the imaginary argument .
the quantity @xmath109 is a part of the scalar electric - dipole ( @xmath110 ) or electric - quadrupole ( @xmath111 ) dynamic polarizability of the state @xmath112 , in which the sum over the intermediate states @xmath113 is restricted to the states with fixed total angular momentum @xmath114 : @xmath115 here , @xmath116 stands for all quantum numbers of the intermediate states except @xmath117 .
the correction @xmath106 to the @xmath118 term in eq.([x_j ] ) is due to a downward @xmath119 transition and is given by the following expression : @xmath120 where @xmath121 the expression for the @xmath122 coefficient is substantially more complicated , so it is discussed in the appendix .
all calculations were carried out by two methods which allows us to estimate the accuracy of the final values .
the first method combines configuration interaction ( ci ) with many - body perturbation theory ( mbpt ) @xcite . in the second method , which is more accurate ,
ci is combined with the coupled - cluster all - order approach ( ci+all - order ) that treats both core and valence correlation to all orders @xcite . in both cases , we start from a solution of the dirac - fock ( df ) equations for the appropriate states of the individual atoms , @xmath123 where @xmath124 is the relativistic df hamiltonian dzuflakoz96b , safkozjoh09 and @xmath125 and @xmath126 are single - electron wave functions and energies .
the calculation was performed in the v@xmath127 approximation , i.e , the self - consistent procedure was done for the [ @xmath128 closed core .
the b - spline basis set , consisting of @xmath129 orbitals for each of partial wave with @xmath130 , was formed in a spherical cavity with radius 60 a.u .
the ci space is effectively complete .
it includes the following orbitals : @xmath131 , @xmath132 , @xmath133 , @xmath134 , and @xmath135 .
the wave functions and the low - lying energy levels are determined by solving the multiparticle relativistic equation for two valence electrons kottup87 , @xmath136 the effective hamiltonian is defined as @xmath137 where @xmath138 is the hamiltonian in the frozen - core approximation .
the energy - dependent operator @xmath139 which takes into account virtual core excitations is constructed using the second - order perturbation theory in the ci+mbpt method @xcite and using linearized coupled - cluster single - double method in the ci+all - order approach @xcite .
@xmath140 in the pure ci approach .
construction of the effective hamiltonian in the ci+mbpt and ci+all - order approximations is described in detail in refs .
the contribution of the breit interaction is negligible at the present level of accuracy and was omitted .
the dynamic polarizability of the @xmath141-pole operator @xmath142 at imaginary argument is calculated as the sum of three contributions : valence , ionic core , and @xmath143 .
the @xmath143 term subtracts out the ionic core terms which are forbidden by the pauli principle
. then @xmath144where both the core and @xmath143 parts are included in @xmath145 .
the valence part of the dynamic polarizability , @xmath146 , of an atomic state @xmath147 is determined by solving the inhomogeneous equation in the valence space .
if we introduce the wave function of intermediate states @xmath148 as @xmath149where `` re '' means the real part , then @xmath150 is given by @xmath151here , @xmath152 is the zeroth component of the @xmath142 tensor .
we include random - phase approximation ( rpa ) corrections to the @xmath141-pole operator @xmath153 .
the eqs .
( [ delphi ] ) and ( [ alphav ] ) can also be used to find @xmath154 , i.e , the part of the valence polarizability , where summation goes over only the intermediate states with fixed total angular momentum @xmath107 .
we refer the reader to ref .
@xcite for further details of this approach .
the core and @xmath143 contributions to multipole polarizabilities are evaluated in the single - electron relativistic rpa approximation . the small @xmath155 term is calculated by adding @xmath143 contributions from the individual electrons , i.e. , @xmath156 and @xmath157 .
a special consideration is required when we need to find the core contribution to @xmath158 of a state @xmath112 .
if we disregard possible excitations of the core electrons to the occupied valence shells , the valence and core subsystems can be considered as independent .
then , the total angular momenta @xmath159 and @xmath160 of the states @xmath112 and @xmath161 , respectively , can be represented as the sum of the valence and core parts @xmath162 .
[ ht ] [ cols=">,>,^,^,>,>,^,^,>,^,^ " , ] @xmath163@xcite , theory .
+ @xmath164@xcite , theory .
+ @xmath165@xcite , based on experiment .
+ @xmath166@xcite , theory .
+ @xmath167@xcite , experiment .
+ @xmath168@xcite , experiment ; the error includes only uncertainty of the fit .
+ @xmath169@xcite , experiment .
the calculation of the @xmath3 and @xmath6 coefficients for the @xmath93 dimer was carried out according to the expressions given by eqs . ( [ c6])-([x_j ] ) and in the appendix [ app ] . a breakdown of the contributions to the @xmath170 coefficient for yb
yb @xmath171 dimer is given in table [ taba ] .
we list the quantities @xmath102 and coefficients @xmath172 given by eqs .
( [ eq1 ] ) and ( [ x_j ] ) for allowed @xmath173 .
the @xmath106 term is presented separately in the second row to illustrate the magnitude of this contribution .
it is relatively small , 4% of the total for @xmath174 and 1% for @xmath175 .
it is included in the @xmath118 value given in the table .
we note that the @xmath176 coefficient do not depend on @xmath177 symmetry .
the ci+mbpt and ci+all - order values for @xmath102 are given in columns labeled `` mbpt '' and `` all '' .
the relative differences between these values , which give an estimate of the higher - order contributions , are listed in the column labeled `` ho '' .
we find that the higher orders contribute with a different sign to @xmath178 and @xmath179 . a breakdown of the contributions to the @xmath180 coefficients for yb
yb @xmath93 dimer is given in table [ tabb ] .
we list the quantities @xmath181 and coefficients @xmath182 ( the analytical expressions for them are given in the appendix [ app ] ) . the @xmath183 and @xmath184 terms are given separately in the first and fifth rows ; they are included in the @xmath185 and @xmath186 contributions , respectively . for calculation of @xmath183 we used the values @xmath187 a.u . and the static @xmath9 polarizability @xmath188 a.u .
obtained in the ci+all - order approximation .
the coefficients @xmath189 and @xmath190 contain @xmath191 , therefore their sign is different for gerade and ungerade symmetry resulting in slightly different values for @xmath192 and @xmath193 . in table
[ tabb ] , the @xmath194 sign corresponds to the ungerade / gerade symmetry , respectively .
the ci+all - order values are given for @xmath195 and @xmath6 ; the relative differences of the ci+all - order and ci+mbpt values are given in column labeled `` ho '' in % .
our final results for polarizabilities and the van der waals @xmath3 and @xmath6 coefficients are summarized in table [ c_68 ] .
the @xmath0 , @xmath1 , and @xmath2 electric - dipole , @xmath196 , and electric - quadrupole , @xmath197 , static polarizabilities in the ci+mbpt and ci+all - order approximations are listed in a.u .. for the @xmath198 state the scalar parts of the polarizabilities are presented .
the values of @xmath199 and @xmath200 coefficients for the @xmath86 dimers in the ci+mbpt and ci+all - order approximations are listed in the second part of the table .
the ( rounded ) ci+all - order values are taken as final .
the relative contribution of the higher - order corrections is estimated as the difference of the ci+all - order and ci+mbpt results , it is listed in column labeled `` ho '' in percent .
we compare frequency - dependent polarizabilities calculated in the ci+mbpt and ci+all - order approximations for all @xmath201 used in our finite grid to estimate the uncertainties of the @xmath3 and @xmath6 coefficients .
we find that the difference between the ci+all - order and ci+mbpt frequency - dependent polarizability values is largest for @xmath202 and decreases significantly with increasing @xmath201 .
this is reasonable because for large @xmath201 the main contribution to the polarizability comes from its core part . but
the core parts are the same for both ci+all - order and ci+mbpt approaches .
therefore , the fractional uncertainty @xmath203 ( @xmath89 ) may be expressed via fractional uncertainties in the static multipole polarizabilities of the atoms @xmath68 and @xmath69 @xcite , @xmath204 the absolute uncertainties induced in @xmath205 and @xmath206 ( @xmath70 ) are given by @xmath207 the polarizabilities and their absolute uncertainties are presented in table [ c_68 ] .
the uncertainties of the electric - dipole @xmath9 and @xmath208 polarizabilities were discussed in detail in ref .
@xcite ; the uncertainty of the @xmath208 polarizability was determined to be 3.4% .
table [ tabe ] illustrates that the accuracy of calculation of the @xmath208 and @xmath198 energy levels is practically the same ( @xmath209 at the ci+all - order stage ) .
we use the same method of solving the inhomogeneous equation to determine both the @xmath208 and @xmath198 polarizabilities .
the main contributions to these polarizabilities are also very similar . based on these arguments , we assume that the uncertainty of the scalar part of the @xmath198 polarizability can be estimated at the level of 3.5% . our estimates of the uncertainties of the electric - quadrupole polarizabilities are based on the differences between the ci+mbpt and ci+all - order values . besides that we take into account that in all cases the dominant contribution comes from the low - lying state which energies we reproduce well ( see table [ tabe ] ) .
based on the size of the higher - order correction , we assign the uncertainties 3 - 4% to these polarizabilities .
these results , as well as the final ( recommended ) values of the polarizabilities , are presented in table [ c_68 ] ( see also ref .
@xcite ) . using eqs .
( [ c_ab_l ] ) and ( [ c_ab ] ) we estimated the fractional uncertainties of the @xmath3 coefficient for the @xmath210 dimers at the level of 44.5% .
the uncertainty of the @xmath211 coefficient is 3.2% and the uncertainties of the @xmath212 coefficients are @xmath213 4.5% .
the difference of the ci+all - order and ci+mbpt values ( 4.5% ) is taken as an uncertainty for the @xmath214 coefficient .
to conclude , we evaluated the electric - dipole and electric - quadrupole static and dynamic polarizabilities for the @xmath0 , @xmath1 , and @xmath2 states and estimated their uncertainties .
the @xmath3 and @xmath6 coefficients are evaluated for the yb - yb dimers . the uncertainties of our calculations of the van der waals coefficients do not exceed 5% .
our result @xmath215 for the @xmath7 dimer is in excellent agreement with the experimental value @xmath216
the quantities calculated in this work allow future benchmark tests of molecular theory and experiment .
most of these quantities are determined for the first time .
methodology developed in this work can be used to evaluate properties of other dimers with excited atoms that have a strong decay channel .
we thank p. julienne for helpful discussions .
this research was performed under the sponsorship of the u.s .
department of commerce , national institute of standards and technology , and was supported by the national science foundation under physics frontiers center grant no .
phy-0822671 and by the office of naval research .
the work of s.g.p . was supported in part by us nsf grant no .
phy-1212442 and rfbr grant no . 11 - 02 - 00943 .
the work of a.d . was supported in part by the us nsf grant no . phy-1212482 .
following formalism of section [ genform ] , the @xmath6 coefficient may be expressed as : @xmath217 which can be further reduced to : @xmath218 where @xmath219 | we evaluate the electric - dipole and electric - quadrupole static and dynamic polarizabilities for the @xmath0 , @xmath1 , and @xmath2 states and estimate their uncertainties .
a methodology is developed for an accurate evaluation of the van der waals coefficients of dimers involving excited state atoms with strong decay channel to the ground state .
this method is used for evaluation of the long range interaction coefficients of particular experimental interest , including the @xmath3 coefficients for the yb
yb @xmath4 and @xmath5 dimers and @xmath6 coefficients for the @xmath7 and @xmath8 dimers . |
tensor models were initially introduced as models for quantum gravity in @xmath3 dimensions @xcite , by generalizing the matrix models @xcite , which are known to consistently describe the @xmath4 dimensional quantum gravity .
the main idea of the tensor models was that the feynman diagrams obtained via perturbative treatment of the tensor models could be identified with the dual diagrams of simplicial manifolds for @xmath3 dimensions , the summation of which would evaluate in a continuum limit the geometric sum in euclidean quantum gravity .
however , these original tensor models were unsuccessful due to singular or un - oriented feynman diagrams @xcite , and also it was not clear how to take continuum limits , because of the absence of @xmath5 expansions denotes the dimension of the vector space associated to the indices of vectors , matrices and tensors in this paper .
namely , an index runs through @xmath6 .
] , which played essential roles in taking the continuum limits of the matrix models .
this situation has been drastically changed by the advent of the colored tensor models @xcite .
the colored tensor models have a one - to - one correspondence between the feynman diagrams and pseudo - simplicial manifolds , and also have @xmath5 expansions , the leading terms of which are composed of melonic diagrams @xcite .
the melonic diagrams correspond to simplicial manifolds with spherical topology , but are geometrically singular @xcite .
therefore it seems important to modify the leading order or to include higher order @xmath5 corrections to obtain continuum limits of physical interests as in the matrix models .
a number of efforts are currently being made in these directions @xcite .
the colored tensor models have also stimulated the developments of the renormalization group procedure of group field theories @xcite , which are tensor models with group - valued indices @xcite and are studied extensively in the context of the loop quantum gravity @xcite .
the developments above of tensor models concern the euclidean case . in quantum gravity with geometric fluctuations , it would be unclear whether the time - like direction can be treated in the same way as the other space - like directions , which is the standard procedure with wick rotation in field theories on flat space - times .
in fact , in the dynamical triangulation model of quantum gravity , it has been shown that simplicial complexes which can be regarded as smooth space - times in global scale dominate in causal dynamical triangulation @xcite , while dynamical triangulation , the original euclidean model , does not seem successful in this respect .-
fields , the authors in @xcite found a promise of a phase transition higher than first order , which , however , is in conflict with the result in @xcite . ] inspired by this fact , one of the present authors proposed hamilton formalism of tensor models , dubbed the canonical tensor model in short @xcite .
this model has a canonical conjugate pair of three - index tensors as dynamical variables , and is formulated as a totally constrained system with first - class constraints , the algebra of which resembles that of the adm formalism @xcite of general relativity .
a number of intriguing results have been obtained so far in this direction : there exists a formal continuum limit in which the first - class constraint algebra agrees with that of the adm formalism @xcite , the model is unique under some physically reasonable assumptions @xcite , the constraints can consistently be quantized to form a first - class quantized constraint algebra with no anomaly @xcite , locality is favored at least for @xmath7 in the physical wave - function satisfying the quantized constraints @xcite , the @xmath8 case classically agrees with the mini - superspace approximation of general relativity @xcite , and the model with @xmath7 has intimate relations with ising model on random networks@xcite .
while the above results suggest that the canonical tensor model can be an interesting model of quantum gravity , a major question on the validity of the model would be whether its quantum dynamics can produce an object like space - time .
this question may be answered by studying the physical states ( or physical wave - functions in explicit representations ) which satisfy the quantized first - class constraints of the canonical tensor model .
this is rather a difficult task due to the non - linear character of the quantized constraints and so far has been solved only in the case of @xmath7 @xcite . on the other hand ,
the previous paper @xcite implies the intimate relations between the canonical tensor model and statistical systems on random networks@xcite .
in fact , the present paper successfully constructs physical wave - functions satisfying the quantized constraints , general for arbitrary integer @xmath1 , in terms of scale free integration of `` grand - type '' partition functions of statistical systems on random networks ( or rather random tensor networks as we will see ) .
we also obtain other physical wave - functions either by generalizing them for any @xmath1 or by explicitly solving the differential equations representing the constraints for @xmath0 .
the following two properties of the solutions suggest physically interesting directions of future study .
one is that the wave - functions can have peaks at configurations where symmetries are enhanced .
this is explicitly observed for @xmath2 , and we expect this to be true also for higher @xmath1 . in future study , this feature may provide a clue as to why our universe is isotropic and homogenous at large scales .
we also find that the physical wave - functions satisfying the constraints for @xmath9 with a cosmological constant can be obtained by restricting the domains of those for @xmath10 with no cosmological constant .
this suggests that , in the canonical tensor model , a cosmological constant is not an input parameter , but rather a part of dynamics .
this would provide an interesting clue to the problem of the cosmological constant in future study .
this paper is organized as follows . in section [ introduction ] , we recapitulate the canonical tensor model and its quantized constraints with a cosmological constant . a classification of physical states in terms of their kinematical and dynamical natures is introduced . in section [ n2 ]
, we obtain the physical wave - functions satisfying the quantized constraints for @xmath7 generally with a cosmological constant by explicitly writing down and solving the partial differential equations . in section [ n3 ] ,
we obtain physical wave - functions on an @xmath11 symmetric subspace of the configuration space in the case of @xmath2 generally with a cosmological constant .
we find that the wave - functions can have peaks at configurations with enhanced symmetries . in section [ sec : ransol ] , we construct physical wave - functions for general @xmath1 with no cosmological constant in terms of scale free integration of partition functions of statistical systems on random networks . in section [ sec : simpsol ] , we obtain physical wave - functions with simpler expressions by generalizing the result in section [ sec : ransol ] . in section [ sec : explicitexample ] , we explicitly evaluate the wave - functions in section [ sec : simpsol ] for @xmath12 to check the validity of the expressions . in section [ sec : solprep ] and [ sec : solmatten ] , we work in a momentum representation and obtain other physical wave - functions for the case of no cosmological constant and general @xmath1 .
we also obtain other types of physical wave - functions in terms of integration over matrix and tensor variables . in section [ sec : wdwcosmo ] , we obtain physical wave - functions for general @xmath1 with a cosmological constant by generalizing the results in section [ sec : solprep ] .
we find that these wave - functions for @xmath9 are just obtained by restricting the domains of the wave - functions for @xmath10 with no cosmological constant .
a general theorem of this aspect is given in section [ sec : generalcosmo ] .
finally , section [ discussions ] is devoted to summary and discussions .
the canonical tensor model has been introduced as a theory of dynamical fuzzy spaces @xcite , aiming to construct a quantum theory of gravity .
first of all , the fuzzy space is an extended notion of space defined by a set of functions , @xmath13 , and the product of such functions : @xmath14 where @xmath15 is a three - index tensor characterizing the fuzzy space .
for instance , in the case of ordinary @xmath16-dimensional euclidean space , ( [ fuzzy ] ) is given by @xmath17 here @xmath18 can be expressed by a @xmath16-dimensional delta function , @xmath19 where @xmath20 .
this implies that points in euclidean space correspond to functions , @xmath18 s .
as the three - index tensor is a delta function , distant points ( functions ) are independent .
one can extend this notion of localized points to non - local `` fuzzy '' ones using a non - trivial three - index tensor , @xmath15 .
therefore , in general a fuzzy space is non - local and can be of any dimension . in the canonical tensor model , in order to define the physics controlling the dynamics , two external conditions are imposed @xcite .
these are the reality conditions given by @xmath21 where @xmath22 means complex conjugation , and the trace - like property of the inner product , @xmath23 where the inner product , @xmath24 , has been chosen to be real , symmetric and bilinear . if one initially sets the inner product to be positive - definite , one can choose an orthonormal basis : @xmath25 by a real linear transformation preserving the two conditions stated in ( [ reality ] ) and ( [ trace ] ) .
this choice of orthonormal basis in ( [ inner ] ) allows to rewrite ( [ fuzzy ] ) in the following form : @xmath26 using the three - index tensor @xmath27 from ( [ tensor ] ) , it is possible to extract the dynamics of the fuzzy space under the imposed conditions stated in ( [ reality ] ) and ( [ trace ] ) .
the inner product in ( [ inner ] ) is @xmath28 invariant , which appears in the transformations of @xmath29 as @xmath30 this symmetry serves as a kinematical symmetry of the canonical tensor model .
the two conditions , ( [ reality ] ) and ( [ trace ] ) , lead to the generalized hermiticity condition of the three - index tensor : @xmath31 the canonical tensor model stands on the position such that space - time would be a time evolution of the dynamical fuzzy space satisfying ( [ reality ] ) and ( [ trace ] ) , or equivalently the generalized hermiticity condition , ( [ hermite ] ) .
this can be realized in analogy with the adm formalism of general relativity @xcite . in the adm formalism of general relativity ,
one considers the @xmath32-dimensional space - time as a time propagation ( history ) of the @xmath33-dimensional spatial hyper - surface , or in other words , parametrizes the metric in terms of non - dynamical fields , @xmath1 , @xmath34 , and dynamical spatial metric , @xmath35 : @xmath36 then one converts the einstein - hilbert action into the hamiltonian using the spatial metric , @xmath35 , and its conjugate momentum , @xmath37 , as phase - space variables . as a result of general covariance , the hamiltonian can be written as a linear combination of constraints : @xmath38 , \label{admh}\ ] ] where @xmath39 and @xmath40 are the generators of temporal and spatial diffeomorphism , respectively .
they are constraints , classified as the first class following dirac s theory of constrained systems .
in fact , the constrains form a closed algebra .
the canonical tensor model has been constructed in the same spirt of the adm formalism , _
i.e. _ , considering the three - index tensor , @xmath27 , and its conjugate momentum , @xmath41 , as dynamical phase - space variables , with the model being introduced as a totally constrained hamiltonian system . one can construct the generators corresponding to the @xmath28-kinematical symmetry , ( [ on ] ) , playing a similar role as the spatial diffeomorphism in general relativity , and remarkably one can uniquely determine the analogue of the temporal diffeomorphism of general relativity , under reasonable assumptions requiring closed constraint algebra , cubic terms at most , invariance under the time - reversal symmetry and connectivity @xcite .
in this paper we are interested in quantizing the canonical tensor model started in @xcite . in the quantum version ,
one replaces the dynamical variable @xmath27 and its conjugate by corresponding operators @xmath42 and its conjugate partner , respectively .
this will be explained later in more detail .
besides , we especially consider the minimal version of the canonical tensor model dubbed _ minimal model _
@xcite , defined by a canonical conjugate pair of real and symmetric tensors , @xmath43 , @xmath44 which satisfy the following commutation relations : @xmath45 = \frac{i}{6 } \sum_{\sigma } \delta_{a\sigma ( d ) } \delta_{b \sigma ( e ) } \delta_{c \sigma ( f ) } , \ \ [ \hm_{abc } , \hm_{def } ] = [ \hp_{abc } , \hp_{def } ] = 0 , \label{commutator}\ ] ] where the summation over @xmath46 implies that over all the permutations of @xmath47 .
as the minimal model has less degrees of freedom , its quantization is easier to perform and the algebra is more tractable compared to the non - minimal version of the canonical tensor model .
however the minimal model is still complicated enough to capture the non - trivial physics of the system .
it therefore becomes a good theoretical laboratory to gain expertise and maturity in the subject ( see , say @xcite ) .
the quantum dynamics of the minimal model can be described through the hamiltonian operator @xcite , @xmath48 } \hat{\mathcal{j}}_{[ab ] } , \label{qh}\ ] ] where @xmath49 and @xmath50}$ ] are lagrange multipliers , and @xmath51 } = \frac{1}{4 } \left ( \hat{p}_{acd } \hat{m}_{bcd } - \hat{p}_{bcd } \hat{m}_{acd } \right ) \label{hj}.\ ] ] here summing over repeated indices is implied , while the symbol @xmath52 $ ] refers to anti - symmetry of indices such that @xmath53}=-\hj_{[ba]}$ ] .
ingredients appearing in ( [ qh ] ) are explained in order : @xmath53}$ ] serve as the generator of the @xmath28-kinematical symmetry , ( [ on ] ) , and @xmath54 is the generator of the symmetry analogous to the temporal diffeomorphism in general relativity ; they are first class constraints and form a closed constraint algebra : @xmath55 = \frac{i}{6}\hj \left ( [ \hat{\xi^1 } , \hat{\xi^2 } ] + 2 \lambda [ \xi^1 , \xi^2 ] \right ) , \notag \\ & [ \hj ( \eta ) , \hh ( \xi ) ] = \frac{i}{6 } \hh \left(\eta \xi \right ) , \label{eq : constraintalg}\\ & [ \hj ( \eta^1 ) , \hj ( \eta^2 ) ] = \frac{i}{6 } \hj \left ( [ \eta^1,\eta^2 ] \right ) , \notag\ ] ] where @xmath56 , @xmath57}\hj_{[ab]}$ ] and @xmath58 ; @xmath59 $ ] denotes the matrix commutator for @xmath60 , and @xmath61_{[ab]}=\xi^1_a \xi^2_b-\xi^2_a \xi^1_b$ ] .
note that this is not a lie algebra with structure constants , but has non - linear structures as on the right - hand side of the first line .
following the standard nomenclature , we call @xmath53}$ ] and @xmath54 _ momentum _ and _ hamiltonian constraints _ , respectively .
@xmath62 is a real constant introduced by the operator ordering and has been fixed by imposing the hermiticity of @xmath54 , while @xmath63 is a real undetermined constant which we call _ cosmological constant_. the last physical naming can be justified by its role in comparison with the mini - superspace approximation of general relativity @xcite .
it is worth to recall that , in the case of hermite tensors , the cosmological constant term is prohibited by the consistency of the constraint algebra @xcite , while this term is allowed for the minimal case as above . in the case
when @xmath64 , one can consistently incorporate the dilation generator to the hamiltonian operator as @xmath65 , where @xmath66 is the lagrange multiplier and @xmath67 in operator form is given by @xcite , @xmath68 here @xmath69 is a constant introduced by the operator ordering , and can be fixed as @xmath70 by imposing the hermiticity of @xmath67 .
the constraint algebra remains closed even with the inclusion of the dilation generator , as can be seen for @xmath64 from the following , @xmath71 = \frac{i}{6}\hh ( \xi ) , \notag \\ & [ \hd , \hj(\eta ) ] = 0.\ ] ] however , we will not use this dilation generator as a genuine generator but merely as a mathematical tool for simplifying the process to find a solution to the constraint equations introduced below . in the following sections
, we will find the physical states satisfying the constraints , @xmath72 } \psi = 0 .
\label{wdw}\ ] ] note that here we do not impose the @xmath73 constraint to define the physical states . for later convenience ,
we introduce the @xmath29-representation of @xmath74 and the @xmath75-representation of @xmath76 as follows : @xmath77 where @xmath29 and @xmath75 are the eigenvalues of @xmath78 and @xmath79 , respectively , and @xmath80 are rescaled partial differentials satisfying @xmath81 finally , let us introduce a classification of the physical states in two types in the case with no cosmological constant , @xmath64 . in this case , from , the hamiltonian constraints can be rewritten in a form , @xmath82 where @xmath83 with the round brackets of @xmath84 symbolically representing the symmetric feature of the two indices .
then , the solutions to the constraints can be classified into the following two types , @xmath85}\psi=0 , \\ & \hbox{dynamical : } \hat { \cal h}_{a}\psi=\hat { \cal j}_{[ab]}\psi=0,\ { } ^\exists \hat { \cal j}_{(ab)}\psi\neq 0,\ ] ] the naming of
kinematical " comes from the fact that the operators , @xmath86 and @xmath87}$ ] , form a @xmath88 lie algebra .
this is a linear lie algebra with structure constants , and such kinematical physical states would reflect only the kinematical characters rather than the dynamics of the canonical tensor model .
a naive expectation is that physically interesting dynamics is caused by the non - linear features of the constraint algebra with structure functions , as in general relativity . from this viewpoint ,
the dynamical states would be of more importance . from
, a necessary condition for a kinematical state , which is a convenient criterion in subsequent analysis , is given by @xmath89 \psi=0.\ ] ] in the @xmath75 and @xmath29 representations , this criterion is respectively represented as @xmath90 \psi(m)=0 .
\label{eq : criteriam}\ ] ]
in this section we will consider the case of @xmath7 tensor model both with and without a cosmological constant . in either case , we will explicitly solve the constraint equations ( [ wdw ] ) to find the physical wave - function . in this subsection
we consider the @xmath7 canonical tensor model with no cosmological constant , thereby setting @xmath91 . in this model , in the @xmath75-representation , there are @xmath32 independent variables , @xmath92 . on the other hand ,
the constraint equations give only three independent first order partial differential equations : @xmath93 } \psi ( p ) = 0 \ , .
\label{wdwn2}\ ] ] this implies that the most general solution to ( [ wdwn2 ] ) can be written with an arbitrary function of a single variable . in order to find this ,
we first consider a solution on a two - dimensional subspace , and then write it in an @xmath94-invariant form to extend it on the whole space .
we start by introducing a @xmath95-dimensional subspace ( or a gauge choice with respect to @xmath96}$ ] and @xmath97 ) in the @xmath75-representation : @xmath98 where @xmath99 are the variables parameterizing the subspace . in this @xmath95-dimensional subspace ,
the constraint equations ( [ wdwn2 ] ) are given by @xmath100 \psi = 0 , \label{n2h1 } \\
\hh_2 \psi & \propto \left [ x_1 ( 1 + 2x_1 ) \frac{\partial}{\partial p_{112 } } + 3x_1x_2 \frac{\partial}{\partial x_1 } + 3(x_1{}^2 + x_2{}^2 ) \frac{\partial}{\partial x_2 } + 5x_2 \right ] \psi = 0 , \label{n2h2 } \\ \hj_{[12 ] } \psi & \propto \left [ ( 1 - 2x_1)\frac{\partial}{\partial p_{112 } } -x_2 \frac{\partial}{\partial x_1 } + 3x_1 \frac{\partial}{\partial x_2 } \right ] \psi = 0 \label{n2j } \ , .\ ] ] by using ( [ n2j ] ) , we remove @xmath101 from ( [ n2h2 ] ) , which is a direction straying away from the subspace ( [ n2sub ] ) .
this gives us the following equation , @xmath102 \psi = 0 . \label{n2h2j}\ ] ] in fact , a solution to ( [ n2h2j ] ) has been found in @xcite : canonical tensor model incorporating also the dilation constraint .
however , ( [ n2withd ] ) of course satisfies ( [ n2h2j ] ) in the present case . ]
@xmath103 then , the general solution to ( [ n2h2j ] ) is given by @xmath104 where @xmath105 satisfies the homogeneous part of ( [ n2h2j ] ) , _ i.e. _ , @xmath106 f=0.\ ] ] this implies that @xmath107 is constant along the _ characteristics _
, @xmath108 this can be solved to obtain , @xmath109 where @xmath110 is a constant . therefore , we get the following result , @xmath111 where @xmath112 is an arbitrary function of @xmath113 . this solution can be written in an @xmath94-invariant form .
as demonstrated in @xcite , on the @xmath95-dimensional subspace ( [ n2sub ] ) , one can explicitly check that @xmath114 where @xmath115 .
therefore , one can extend the solution above to an @xmath94-invariant form : @xmath116 in fact , one can explicitly check that ( [ n2sol ] ) solves ( [ wdwn2 ] ) on the whole space . since the solution ( [ n2sol ] ) contains an arbitrary function of a single variable , @xmath112 , it should be the most general solution to ( [ wdwn2 ] ) .
let us classify the general solution into the kinematical and dynamical solutions discussed in the last of section [ introduction ] .
the necessary criterion for a kinematical solution requires that a wave - function be a homogenous function of @xmath75 with degree @xmath117 .
this determines @xmath118 , and therefore , @xmath119 in fact , one can explicitly check that , not only the necessary criterion , but also the full conditions for a kinematical state , @xmath120 , are satisfied by .
then , the other wave - functions than are the dynamical ones .
this in turn implies that the potential singularities at @xmath121 of the general wave - function are purely of dynamical origin .
as was argued in @xcite , the configurations satisfying @xmath121 are those of maximal locality .
this would mean that the non - linear character of the constraints is essentially important in the emergence of locality in the canonical tensor model .
the characteristics of the configurations satisfying @xmath122 , which are of kinematical origin as shown above , can be discussed in terms of the fuzzy space interpretation presented in section [ introduction ] .
let us consider the following condition : @xmath123 where @xmath124 is a non - vanishing vector .
then , in the gauge , one can easily show that such a non - vanishing @xmath124 exists , if and only if @xmath125 , which exactly corresponds to @xmath122 .
in the fuzzy space interpretation of the canonical tensor model explained in section [ introduction ] , the `` points '' in a fuzzy space , @xmath126 s , are assumed to form an algebra or @xmath75 defines the fuzzy space algebra . at present , we have no good argument to determine which choice is more proper . ] , @xmath127 thus the condition implies that a point " , @xmath128 , satisfies @xmath129 therefore , @xmath122 has the meaning that there exists a `` point '' , @xmath107 , with the property of a grassmann number .
we will see in section [ n3without ] that configurations with the same nature appear as potential singularities also in the case of @xmath2 .
we now consider the @xmath7 tensor model with a cosmological constant .
since the corresponding constraint equations are more involved and finding the most general solution does not seem straightforward on first sight , we will first solve them on a subspace invariant under an @xmath130 transformation .
we will then extend the solution over the whole space by finding an @xmath94 invariant expression .
the reason why we consider in this subsection such a symmetric subspace rather than is that the equations are simplified because of the symmetry and a similar procedure as below can also be applied to the case of @xmath2 as in section [ n3 ] .
the subspace we consider is parameterized by @xmath131 which are the fixed points of the @xmath130 transformation permuting the index set , @xmath132 .
the method of characteristics employed in section [ n2without ] reduces the problem of solving the set of the first - order partial differential equations representing the constraints to solving the ordinary first - order differential equations along the flows generated by the constraints .
an infinitesimal variation of @xmath79 generated by the constraints is given by @xmath133 } \hj_{[de ] } \right ] \
, , \label{n2vari}\ ] ] where @xmath134 and @xmath135 are infinitesimal parameters .
one can easily show that the infinitesimal variation ( [ n2vari ] ) goes out of the subspace ( [ n2x1x2 ] ) , unless we require @xmath136}=0 \ , .\ ] ] this requirement in turn determines the constraint equations on the @xmath130 subspace as @xmath137 where @xmath138 is an wave - function . by noting that @xmath139 ( [ n2wdwlam ] ) becomes @xmath140 \psi_{\lambda } = 0 .
\label{eq : partdifn2}\ ] ] from the result of section [ n2without ] , we know that the solution for @xmath64 is given by with @xmath141 ^ 2,\ ] ] where @xmath142 have been obtained by putting to and , respectively .
in fact , @xmath143 satisfy the following peculiar properties , @xmath144 where @xmath145 with @xmath146 defined by the derivative part of , @xmath147 this means that with @xmath64 can be rewritten as a partial differential equation with variables @xmath148 rather than @xmath149 , because , from the leibniz rule , @xmath150 \psi(y)&=\left [ \left({\cal o}_0 a \right)\frac{\partial}{\partial a } + \left({\cal o}_0 b \right)\frac{\partial}{\partial b}+10 \left ( y_1 + y_2 \right ) \right ] \psi(a , b ) \cr & = ( y_1+y_2)\left [ 12 a \frac{\partial}{\partial a } + 16 b \frac{\partial}{\partial b } + 10 \right ] \psi(a , b)=0.\ ] ] when @xmath148 are regarded as variables , the expression is indeed the most general solution to the simple partial differential equation in the last line .
therefore , if we keep the properties and for general @xmath63 , the solution to is simply given by the same expression as with proper replacements of @xmath151 and @xmath152 .
namely , the solution is given by @xmath153 where @xmath112 is an arbitrary function , and @xmath154 must be determined through @xmath155 it does not seem always guaranteed that one can obtain the solutions of @xmath156 to in simple expressions .
but , rather miraculously , we obtain @xmath157 ^ 2 \label{n2b}.\ ] ] the reason why we get such simple expressions even for general @xmath63 is probably related with what will be discussed in section [ sec : generalcosmo ] .
the solution in the whole space can simply be obtained by finding the @xmath94-invariant expressions of @xmath158 : @xmath159 ^ 2 \label{n2bp},\ ] ] where @xmath160 in fact , one can explicitly check this is the solution in the whole space .
in this section we extend our horizons further and explore the @xmath2 tensor model both with and without a cosmological constant . while , in this more complicated model , it is hard to find the general solution to the constraint equations ( [ wdw ] ) , it is possible to find the most general solution in a subspace satisfying an @xmath11 symmetry .
the procedure is basically the same as that employed for @xmath7 in section [ n2with ] .
here we consider the @xmath2 minimal tensor model with no cosmological constant , and find the general solution to the constraint equations on a subspace invariant under an @xmath11 symmetry .
the subspace is parametrized by @xmath161 for any @xmath162 , which are the fixed points of the @xmath11 transformations permuting the index set @xmath163 . as in section [ n2with ]
, one can show that there exists only one linear combination of the constraints which generates a flow along the @xmath11 symmetric subspace , and it is given by @xmath164 on the subspace ( [ n3x1x2x3 ] ) , the constraint ( [ n3const ] ) is expressed as @xmath165 where @xmath166 , and we have used @xmath167 in order to further simplify the process of finding a solution , let us virtually impose the dilation constraint ( [ hd ] ) : @xmath168 where @xmath169 is supposed to be a free parameter for the purpose , rather than the fixed value in . by assuming the two constraints , and , and @xmath170 for the time being
, we can eliminate @xmath171 to obtain a constraint , @xmath172 the method of characteristics applied to this partial differential equation , @xmath173 , reduces the problem to solving @xmath174 where @xmath175 parametrizes the trajectory of the flow .
note that @xmath176 is regarded as a constant in these equations .
the trajectory equation derived from ( [ ds ] ) , @xmath177 tells us that the following combination remains unchanged along the trajectory : @xmath178 this gives a motivation to consider an ansatz : @xmath179 where @xmath180 can be read off from ( [ ds ] ) , and @xmath181 s are numerical constants to be determined . inserting ( [ n3vpsi ] ) into the equations , @xmath182 , one obtains @xmath183 since @xmath184 and @xmath69 can be taken arbitrary to satisfy the single constraint equation , @xmath185 , the general form of @xmath186 is determined to be @xmath187 where @xmath107 is an arbitrary function with two arguments .
it is clear from number counting that ( [ n3solwoc ] ) gives the most general solution to the constraint equation , @xmath188 .
let us study peaks of the wave - function in ( [ n3vpsi ] ) . since we can not uniquely determine the coefficients , @xmath181 , here we consider possible peaks of the wave function ( [ n3vpsi ] ) assuming signs of the coefficients .
firstly , if @xmath189 , then ( [ n3vpsi ] ) diverges when @xmath190 .
this means that the configurations satisfying this equation are enhanced .
in fact , it can be shown that , if @xmath190 , @xmath41 is invariant under an infinitesimal so(@xmath95 ) transformation : @xmath191 where @xmath192 therefore , the configurations with the so(@xmath95 ) symmetry are enhanced in the wave - function . secondly , if @xmath193 , then the wave - function ( [ n3solwc ] ) diverges when @xmath194
this configuration satisfies @xmath195 where @xmath196 this is the same situation discussed for the case of @xmath7 in the last paragraph of section [ n2without ] .
therefore , the condition , @xmath194 , has the meaning that there exists a `` point '' , @xmath197 , with the property of a grassmann number .
a more interesting possibility arises for the configurations which simultaneously satisfy the two conditions above , @xmath198 .
they are the configurations satisfying @xmath199 what is interesting on these points is that , in addition to the so(@xmath95 ) symmetry mentioned above , @xmath41 is invariant under an infinitesimal boost transformation , @xmath200 where @xmath201 therefore , these configurations are invariant under a lorentz group . in the present @xmath2 case , the generators commute with each other , @xmath202=0 $ ] , and may not be of much interest .
however , we expect enhancement of non - abelian lorentz groups to occur for larger @xmath1
. we could not find similar kinematical characterization as above for the other possible peaks at @xmath203 or @xmath204 .
probably , these are more rooted in dynamics , similarly to @xmath121 for @xmath7 as discussed in section [ n2without ] . in this subsection
, we generalize the solution on the @xmath11 symmetric subspace obtained in section [ n3without ] to include the cosmological constant @xmath63 by applying the method employed in section [ n2with ] for the case of @xmath7 . with @xmath63 ,
the constraint is replaced by @xmath205 as in section [ n2with ] , let us take the derivative part of the constraint as @xmath206 the solutions ( [ n3solwoc ] ) for @xmath91 are expressed in terms of the following linear combinations of @xmath207 s : @xmath208 in the same way as the @xmath7 case , one would be able to construct solutions in the case of @xmath209 by considering corrections to these @xmath210 so that they satisfy the requirement corresponding to .
in fact , @xmath211 satisfies @xmath212 the crucial point to note here is that the right - hand side of ( [ n3de1 ] ) is independent of @xmath63 , and therefore @xmath211 meets the requirement without any corrections .
let us next consider @xmath213 . in this case ,
@xmath214 which requires us to make corrections to @xmath213 .
since the dimension of @xmath63 is that of @xmath215 , @xmath213 is not appropriate to consider corrections in perturbation of @xmath63 .
instead , we would be able to consider @xmath216 and its correction as @xmath217 where @xmath218 is a numerical constant to be determined . applying ( [ n3d ] ) on @xmath219 ,
one obtains @xmath220 therefore , if one takes @xmath221 in ( [ e3 ] ) , then @xmath222 the righthand side does not contain @xmath63 , and therefore @xmath219 satisfies the requirement . in a similar manner , we can find suitable combinations of variables as @xmath223 thus , using the new set of variables , @xmath224 , one can generalize ( [ n3solwoc ] ) to @xmath225 where @xmath226 is an arbitrary function .
in fact , one can explicitly check that ( [ n3solwc ] ) satisfies @xmath227 as shown above for @xmath2 , we have again found that , similar to the case of @xmath7 , the inclusion of the cosmological constant can be implemented by making some corrections to the solutions with no cosmological constant .
this suggests that the inclusion of the cosmological constant does not change the essential structure of the theory .
in fact , we will see in section [ sec : generalcosmo ] that the cases with a cosmological constant can be treated by considering the cases with no cosmological constant for increased @xmath1 .
so far we have studied the solutions of the constraint equations of the canonical tensor model for the simple cases of @xmath0 by explicitly writing down and solving the sets of the partial differential equations representing the constraints . while this procedure can work for small @xmath1 s ,
it is obvious that the situation becomes much more complicated as soon as one moves to higher-@xmath1 tensor model , and one quickly notices the limitations of the methodologies used in studying the @xmath0 cases . in the present and following sections , instead of giving general explicit solutions for particular @xmath1 s , we will present some series of solutions valid for general @xmath1 , by using the partition functions of statistical systems on random networks and some variants . in @xcite , it was argued that the @xmath7 canonical tensor model is intimately related to the ising model on random networks of trivalent vertices .
in fact , it has been shown that the phase structure of the ising model on random networks can be derived from the hamiltonian vector flow of the canonical tensor model for @xmath7 , if the flow is regarded as the renormalization group flow of the ising model .
the present and following sections will give further relations between the canonical tensor model and statistical systems on random networks by obtaining the physical wave - functions solving the constraints from the perspective of statistical systems on random networks .
let us start with the following form of the grand - type " partition function of statistical systems on random networks @xcite , @xmath228 where we have used shorthand notations , @xmath229 here @xmath230 and @xmath231 are a numerical variable and a three - index symmetric tensor , respectively .
@xmath232 denotes the domain of integration over @xmath233 s , which can generally take complex values . as discussed in @xcite , to make relations with statistical systems on random networks , @xmath230 is assumed to be a positive real number , and the integration contour , @xmath232 , is taken , for instance among various allowed possibilities , as @xmath234 where @xmath235 run from @xmath236 to @xmath237 on the real axis . by expanding the integrand of in @xmath29 , one obtains an asymptotic expansion in @xmath29 as @xmath238 each @xmath239 gives the partition function of a statistical system on random networks of @xmath240 trivalent vertices @xcite . in this paper
, we do not generally assume the positivity of @xmath230 or the integration contour like , since our interest is in the solutions to the constraint equations rather than such statistical systems . as we will see
shortly , the necessary properties are the well - definedness of the integral in the space of @xmath29 except for possible singularities , and the validity of the partial integrations over @xmath233 s which will be performed in the rest of this paper .
firstly , one can show that the partition function satisfies the momentum constraints ( [ hj ] ) , because @xmath241 } z(k , m,{\cal c}_\phi)&\propto ( m_{acd } d^m_{bcd}-m_{bcd } d^m_{acd})z(k , m,{\cal c}_\phi ) \cr & = \int_{\cal c_\phi } d\phi\ ( m_{acd } \phi_b \phi_c \phi_d -m_{bcd } \phi_a \phi_c \phi_d ) e^{-k \phi^2+m\phi^3 } \cr & = \frac{1}{3 } \int_{\cal c_\phi } d\phi\ e^{-k \phi^2 } \left ( \phi_b d^{\phi}_a
-\phi_a d^\phi_b \right)e^{m\phi^3 } \cr & = -\frac{1}{3 } \int_{\cal c_\phi } d\phi\ e^{m\phi^3 } \left[d^{\phi}_a\left ( \phi_b e^{-k \phi^2}\right ) - d^\phi_b \left(\phi_a e^{-k \phi^2}\right ) \right ] \cr & = 0 \ , , \ ] ] where we have used a short - hand notation , @xmath242 and have assumed the validity of the partial integrations over @xmath233 s .
this is justified , if @xmath243 is a closed curve , or if the integrand damps rapidly enough in the case that @xmath243 extends to infinity . in the rest of this paper , without being explicitly mentioned , we simply assume that all the partial integrations ( not only over @xmath233 s but also over some other variables ) performed in due course be valid .
next , let us consider the hamiltonian constraints . from ,
the corresponding partial differential equations for @xmath64 in the @xmath29-representation are given by @xmath244 as for the first term , we obtain @xmath245 as for the second term of , we obtain @xmath246 because of the derivative term with respect to @xmath230 in the last line of , the partition function itself does not satisfy the hamiltonian constraints .
this can be remedied by introducing a function @xmath247 and the integration over @xmath230 . to see this ,
let us define @xmath248 then @xmath249 e^{-k \phi^2+m\phi^3 } \\ & \hspace{2 cm } = \frac{2}{3 } \int_{{\cal c}_{k,\phi } } dkd\phi\ \phi_a \phi^2 \left ( k g'(k)+\alpha_n g(k ) \right ) e^{-k \phi^2+m\phi^3 } , \ ] ] where we have performed a partial integration over @xmath230 under the assumption of its validity , and @xmath250 therefore the hamiltonian constraints are satisfied , if we take @xmath251 by considering various choices of the integration domain , @xmath252 , one can obtain a number of independent solutions .
physics behind the present solution would be given as follows .
as discussed in the previous paper @xcite , a flow generated by a scale - free @xmath28-invariant linear combination of the hamiltonian constraints of the canonical tensor model could be interpreted as a renormalization group flow of statistical systems on random networks .
corresponding to the ising model . ] a renormalization group procedure , if existed , on random networks would change the number of vertices of networks , that is a natural analogy of a block spin transformation on a regular lattice . therefore , since the parameter @xmath230 controls the relative weights among networks of @xmath240 trivalent vertices as @xmath253 , a renormalization group procedure should generate a flow in @xmath230 .
this would be in accordance with the appearance of a derivative term of @xmath230 in the last line of , and it must be compensated by introducing @xmath247 to make an invariant under the flow . here
let us check the type of the solution @xmath254 .
we obtain @xmath255 here the second term in the last line can not be concluded to be proportional to @xmath256 , and therefore @xmath257 ( see for @xmath86 ) is expected in general .
this means that @xmath254 is a dynamical solution in general .
however , there exists a delicate issue for @xmath258 as follows . by performing similar computations as
, we obtain @xmath259 then one finds that , for @xmath258 , @xmath260 satisfies the necessary condition to be kinematical . for @xmath258
, we can not ignore the possibility that @xmath260 may become kinematical .
for @xmath261 , since the necessary condition is violated , the solution is definitely dynamical .
finally , we would like to comment on the partition function ( [ eq : zkmc ] ) from a mathematical view point .
let us start with a generalization of the airy function : @xmath262 = \int_{\mathcal{c}_{\phi } } d\phi \ e^{-j_a\phi_a+m_{abc}\phi_a\phi_b\phi_c } , \label{generalairy}\ ] ] where @xmath263 is a vector .
( [ generalairy ] ) satisfies the following generalization of airy s differential equation : @xmath264 = 0,\ ] ] which can be shown by partial integrations over @xmath265 .
the generalized airy function can be related to the partition function in the following manner .
firstly , one can remove the linear term in the exponent in ( [ generalairy ] ) by a shift , @xmath266 where @xmath267 is a vector satisfying @xmath268 then there appears a quadratic term in @xmath265 , @xmath269 , as well as a constant term , @xmath270 . the quadratic term can be diagonalized and normalized by a linear transformation , @xmath271 with a complex matrix @xmath272 , if the quadratic term is not singular . then
, ( [ generalairy ] ) can be transformed to a form , @xmath262 = |r| e^ { -j w+m w^3 } \int_{\mathcal{c}'_{\phi } } d\phi \
e^{-k\phi^2+m'\phi^3 } = |r|e^ { -j w+m w^3}\ , z(k , m',{\cal c}'_\phi),\ ] ] where @xmath273 @xmath230 is a number , and @xmath274 denotes the determinant of @xmath272 .
therefore the mathematical properties of the generalized airy functions can be related to those of the partition function .
this observation would be useful in future study .
the solution found in section [ sec : ransol ] has a direct connection with statistical systems on random networks , and would therefore be interesting from physical viewpoints .
on the other hand , if we set aside the physical interpretation , the solution would be simplified by replacing @xmath275 with a function of @xmath276 .
thus , in this section , let us assume a form , @xmath277 where @xmath107 and @xmath278 are functions of the shown arguments , and solve the constraints . @xmath107 will be determined shortly .
let us first check the momentum constraints .
we obtain @xmath279 as for the hamiltonian constraints , we first obtain @xmath280 where @xmath281 .
similarly , @xmath282 therefore , @xmath283 where @xmath284 was defined in .
thus , by putting @xmath285 a solution to the constraint equations can be obtained as @xmath286 where @xmath287 .
one would be able to obtain a number of independent solutions by considering various @xmath232 and @xmath278 .
one can obtain another kind of solutions by determining @xmath278 instead of @xmath107 in . in this case
, we have @xmath288 this vanishes , if @xmath289 where @xmath290 .
the constant term is irrelevant , since it does not produce any dependence on @xmath29 .
thus a non - tivial solution to the constraints is given by @xmath291 various choices of @xmath292 will provide a number of independent solutions . for instance , by considering @xmath293 in , we obtain a similar expression as @xmath239 in . therefore , in the large-@xmath1 limit , one would be able to employ a similar saddle - point method used in the analysis of the thermodynamic limit of statistical systems on random networks in @xcite
. however , there exist at least the following major differences from the statistical systems : the signature of the exponent is opposite , and one can more freely take the integration domain , @xmath232 . therefore , the dynamics would be much different .
we leave this interesting aspect for future analysis . from and , it is obvious that @xmath294 is a solution . in this case , the dimension of @xmath265 is canceled in the integration , since @xmath295 .
therefore , it should be the canonical form of the solution in the case that @xmath232 is a finite domain .
let us compute the scaling dimension of the above solutions .
we easily obtain @xmath296 for all the solutions above , and this coincides with
. then the solutions in this section are dynamical except for the delicate case @xmath258 as discussed in section [ sec : ransol ] .
in this section , we will check the solutions for the simple cases with @xmath258 . for the @xmath8 case , taking @xmath297 in gives @xmath298 here the contour is assumed to be a closed path surrounding the origin .
this is actually the trivial solution to the constraint equation for @xmath8 , @xmath299 to obtain a non - trivial solution , we consider with @xmath300 : @xmath301 which is indeed the other independent solution to . here
the multi - valuedness of the logarithmic function in the integrand is _ not _ problematic : the first derivative , @xmath302 , is a single - valued function , and one can safely justify the partial integration performed in . since @xmath303 ,
this is a dynamical solution . for the @xmath7 case ,
gives @xmath304 let us check if one can really choose the contour , @xmath305 , so that ( [ n2exsol ] ) becomes a non - vanishing meaningful solution . to begin , when choosing the gauge , @xmath306 one finds @xmath307 where @xmath308 s are the three solutions to @xmath309 .
we then implement the @xmath310-integration by choosing the contour , @xmath311 , in such a way as to enclose @xmath312 two times and @xmath313 one time ( see fig.[cont ] ) .
we take such a contour to make it consistent with the multi - valuedness of the integrand caused by the fractional power . :
a closed dashed line winding two times around @xmath312 and one time around @xmath313.,scaledwidth=40.0% ] this choice of contour is in fact democratic : every contour enclosing @xmath314 ( @xmath315 ) two times and @xmath316 @xmath317 one time yield the same result up to a sign .
after the @xmath310-integration , ( [ n2exsol ] ) can be shown to be @xmath318^\frac{1}{3}}. \label{afterphi1}\ ] ] choosing @xmath319 to enclose @xmath320 , one finally finds @xmath321^\frac{1}{3}}\propto \frac{1}{(4 x_1 ^ 3+x_2 ^ 2)^\frac{1}{6}}\propto a(m)^{-\frac{1}{6}},\ ] ] where the last expression is valid in the whole space with @xmath322 defined in .
this can be shown to be a kinematical solution , similarly to .
the method in section [ sec : simpsol ] can also be used to solve the constraint equations in the @xmath75-representation .
let us start with an assumption , @xmath323 where @xmath107 will be determined shortly .
the momentum constraints can be shown to be satisfied by the same reasoning as in section [ sec : simpsol ] . for @xmath64 in ,
the hamiltonian constraints are given by @xmath324 as for the first term , @xmath325 in the last line of , the second term seems to be problematic for the constraint equations to be satisfied .
so , we set @xmath326 to discard it . then , from , we obtain that @xmath327 satisfies @xmath328 the only difference of from is the numerical coefficient of the second term , and it is obvious that @xmath329 satisfies the constraint equations . the scaling dimension can be computed as @xmath330 thus , @xmath331 which agrees with the necessary condition for a kinematical solution .
the scaling dimension will become obvious , if we take @xmath332 as @xmath333 this should be a canonical expression when @xmath232 is a finite domain , since the dimension of @xmath265 is canceled in the integration .
finally , let us check the type of the solution . by partial integrations , one obtains @xmath334 this leads to @xmath335 ,
namely , @xmath336 is a kinematical solution , and might not be so interesting by itself .
however , it will be useful as a component to generate a variety of dynamical solutions in the following sense .
let us suppose that we have two solutions , say @xmath337 , which both satisfy the constraint equations .
then , it is obvious that the product , @xmath338 , with an arbitrary number @xmath339 satisfies .
namely , one can construct one parameter family of solutions to the constraint equations from two independent solutions .
this essentially comes from the fact that the hamiltonian constraints are first - order differential operators in the @xmath75-representation .
now , a number of dynamical solutions in the @xmath75-representation can be obtained through the fourier transform of the solutions in the @xmath29-representation in section [ sec : ransol ] and section [ sec : simpsol ] .
then , the above products of these solutions and @xmath336 will give families of dynamical solutions .
the construction of the solutions in section [ sec : solprep ] can be generalized to that with integration over matrix and tensor variables .
the first two solutions below are kinematical , while the final one is dynamical .
let us consider @xmath340 where we have introduced an integration over a symmetric matrix @xmath341 , in addition to the integration over @xmath265 as before , @xmath342 are functions , and @xmath343 by performing partial integrations as in the previous sections , one can show that @xmath344 where @xmath345 denotes partial derivatives with respect to @xmath341 as @xmath346 then , @xmath347 satisfies the constraint equations . also shows that @xmath348 is a kinematical solution .
a solution with integration over a tensor variable starts with considering @xmath349 where @xmath350 is a symmetric tensor with three indices , and @xmath351 by partial integrations as before , one can show that @xmath352 then @xmath353 can be shown to satisfy the constraint equations .
also shows that @xmath354 is a kinematical solution .
so far , in this and previous sections , we have only constructed kinematical solutions in the @xmath75-representation .
a dynamical solution can be constructed by starting with the following integration over a symmetric matrix @xmath341 , @xmath355 where @xmath356 as for the first term of the hamiltonian constraints , @xmath357 where the underlined indices are supposed to take the average over their permutations , and @xmath358 in a similar way , the scaling dimension can be obtained as @xmath359 these results show that @xmath360 is a solution to the constraint equations , and its scaling dimension is given by @xmath361 by comparing with , @xmath362 is a dynamical solution for @xmath363
. it would be meaningful to see whether @xmath362 is different from the solutions obtained in section [ sec : ransol ] and [ sec : simpsol ] . from ,
we obtain @xmath364 this differs from except for @xmath365 .
therefore , @xmath362 is a new solution at least for @xmath366 .
the canonical tensor model with variables satisfying the generalized hermiticity condition does not allow a cosmological constant term , because such a term violates the consistency of the first - class constraint algebra @xcite .
on the other hand , if we restrict the variables to be real symmetric as in this paper , the algebra remains consistent even after such a cosmological term is introduced as in section [ introduction ] .
the hamiltonian constraint equations with a cosmological constant @xmath63 in the @xmath75-representation are given by @xmath367 where the dependence of the normal ordering term on @xmath1 is explicitly written for later convenience .
the interpretation of @xmath63 as a cosmological constant has been validated by comparing the classical dynamics of the canonical tensor model with @xmath8 and the mini - superspace approximation of general relativity @xcite . in this section
, we will explicitly construct a few series of solutions to by generalizing the solutions obtained in the previous sections .
remarkably , it will be observed that all these solutions for @xmath9 can actually be obtained from those to the constraint equations with _ no _ cosmological constant for @xmath10 by fixing the extra components of @xmath75 . in section [ sec : generalcosmo ] , we will prove a general theorem on this aspect .
let us start with , @xmath368 h(p\phi^3)=0 , \label{eq : startwdw}\ ] ] where @xmath369 is assumed to have the expression on the righthand side of . by performing a partial integration ,
the last term can be computed to be @xmath370 where @xmath371 is a function satisfying @xmath372 now let us assume @xmath373 , \label{eq : hisexp } \\ & f''(x)=b x f(x ) ,
\label{eq : condfx}\ ] ] with numerical constants , @xmath148 . then can further be computed as @xmath374 this concludes @xmath375 q_{{\cal c}_\phi , f , h}(p)=0 .
\label{eq : wdwwithcos}\ ] ] thus , if we consider @xmath376 we obtain @xmath377\psi_{{\cal c}_\phi , f , h}(p)=0.\ ] ] namely , @xmath378 satisfies the constraint equations with a cosmological constant , @xmath379 in fact , the solution to is given by the airy function , @xmath380 = \int_{{\cal c}_z } dz\ \exp \left [ -b^\frac{1}{3 } x z + \frac{z^3}{3 } \right ] .
\label{eq : fairy}\ ] ] then , by putting @xmath381 and into the expression , one obtains an intriguing expression , @xmath382 \cr & = \hbox{const .
} \int _ { { \cal c}_{\phi , z}}d\phi dz \ \exp \left[p\phi^3-\phi^2 z + \frac{a^2}{3 b } z^3 \right ] \cr & = \hbox{const . } \int _ { { \cal c}_{\phi } } \prod_{a=1}^{n+1 } d\phi_a \exp\left[\tilde p\phi^3 \right]\ ] ] with @xmath383 where we have rescaled @xmath384 , and have renamed @xmath385 .
thus , @xmath386 has the form as , and therefore @xmath387 actually satisfies the constraint equations with no cosmological constant for the variables @xmath388 , as same as the wave - function in .
this shows that the cosmological constant can be absorbed into some of the dynamical variables by increasing @xmath1 .
below , we will show that a similar fact holds also for a wider class of solutions .
let us consider the solution , @xmath389 in , and start with , @xmath390 where @xmath391 .
let us divide the index set as @xmath392 and consider a subspace in which @xmath75 s containing the index @xmath113 are fixed as @xmath393 where @xmath394 .
below , @xmath395 will be used for the elements in @xmath396 , and @xmath397 for both @xmath396 and @xmath113 . on the subspace ,
the second term of for @xmath398 becomes @xmath399 on the subspace , the first term of for @xmath400 can be evaluated as @xmath401 h'(p\phi^3 ) , \label{eq : ppdqphi}\ ] ] where it is important to notice that do not contain any conditions on the corresponding derivatives , @xmath402 .
the following two identities hold for : @xmath403 h'(p \phi^3 ) , \cr & 0=\int_{{\cal c}_\phi } d\phi\ d^\phi_z\left ( \phi_i h(p\phi^3)\right ) = \int_{{\cal c}_\phi } d\phi \ \left [ 3 a \phi_i ( \phi_z)^2 + 3 b \phi_i \phi_j^2 \right ] h'(p\phi^3).\ ] ] these identities can be used to delete the terms , @xmath404 and @xmath405 in , to obtain @xmath406 h'(p\phi^3 ) \cr & = \left [ p_{ijk}p_{jlm } d^p_{klm } + \frac{4 b^3}{a } d^p_{ijj } \right ] q_{{\cal c}_\phi , h } ( p)\ ] ] thus the wave - function satisfies the constraint equations with a cosmological constant , @xmath407
in this section , as generalization of section [ sec : wdwcosmo ] , we will prove that a solution to the constraint equations with no cosmological constant for @xmath10 can always generate a solution to the constraint equations with a cosmological constant for @xmath9 by fixing the extra components of @xmath75 as in with @xmath408 .
let us divide the index set as in , and assume that @xmath409 denote the elements in @xmath396 , while @xmath397 both @xmath396 and @xmath113 .
let us assume that a wave - function @xmath410 satisfies the constraint equations with no cosmological constant and the momentum constraints as @xmath411 \psi(p)=0 , \label{eq : wdwpsi}\\ & \left [ p_{acd } d^p_{bcd } - p_{bcd } d^p_{acd } \right ] \psi(p)=0 , \label{eq : oninv}\ ] ] where the dependence of the normal ordering term on @xmath412 is explicitly written .
let us assume for @xmath413 , while @xmath414 are left arbitrary .
then , on the subspace , the second term in for @xmath400 is given by @xmath415 as for the first term in , by putting , we obtain @xmath416 \psi(p ) \cr & = \left [ p_{ijk } p_{jlm } d^p_{klm}+2 p_{ijk } p_{jlz } d^p_{klz } + p_{ijz } p_{jkl}d^p_{zkl}+2 p_{ijz } p_{jkz}d^p_{zkz } \right . \cr & \ \ \ \ \ \ \ \ \left .
+ p_{izj } p_{zkl}d^p_{jkl}+p_{izj } p_{zzz}d^p_{jzz } \right ] \psi(p ) \cr & = \left [ p_{ijk } p_{jlm } d^p_{klm}+3 b p_{ijk } d^p_{jkz}+(2
b ^2+ab ) d^p_{izz}+ b^2 d^p_{ijj } \right ] \psi(p ) .
\label{eq : ppdpqpgen}\ ] ] on the other hand , by putting into for @xmath417 , we obtain @xmath418 \psi(p ) \cr & = \left [ p_{ijk } d^p_{zjk}+2 p_{ijz } d^p_{zjz } - p_{zjk } d^p_{ijk}-p_{zzz } d^p_{izz } \right ] \psi(p ) \cr & = \left [ p_{ijk } d^p_{jkz}+(2 b - a ) d^p_{izz } - b d^p_{ijj}\right ] \psi(p ) .
\label{eq : momrel}\ ] ] by adding multiplied by a free parameter @xmath419 , one finds that the last line of is equivalent to @xmath420 \psi(p).\ ] ] here , by choosing @xmath421 , one can delete the second and third terms to obtain @xmath422 \psi(p ) .
\label{eq : ppdeval}\ ] ] then , from , and , we conclude @xmath423 \psi(p)=0\ ] ] holds on the subspace .
finally , we define @xmath424 then , @xmath425 satisfies the constraint equations , @xmath426 \psi_n(p)=0,\ ] ] with a cosmological constant , @xmath427 two comments are in order .
one is that , to obtain a positive cosmological constant , we have to perform analytic continuation of @xmath428 to pure imaginary values . on the other hand , the explicit general solutions with a cosmological constant for @xmath0 in section
[ n2 ] and [ n3 ] show that @xmath63 appears without such imaginary numbers .
this suggests that there might be another more improved discussion to ignore a cosmological constant than above .
the second is that it is totally unclear whether all the solutions to the constraint equations with a cosmological constant can be obtained from those with no cosmological constant by the present method of increasing @xmath1 .
what we have shown is merely the reverse ; the latter to the former .
therefore , at present , we do not know whether we can totally ignore a cosmological constant in the analysis of the canonical tensor model .
we leave this interesting possibility for future study .
in this paper , we studied exact physical states in the canonical tensor model .
our interest in this paper was to study the canonical tensor model of higher @xmath1 with and without a cosmological constant , and wherever possible to find the exact general solutions to the constraint equations .
after introduction , we presented an outline of the canonical tensor model in section [ introduction ] .
here we described the constraint algebra and the corresponding equations satisfied by the physical states of the canonical tensor model . here
we also introduced the concept of kinematical and dynamical parts of the states , where the latter reflects the non - linear characteristics of the constraints and would be physically more relevant .
we then proceeded to study the canonical tensor model of various @xmath1 .
the achievements of this paper can be summarized into the following three categories .
the first is that we have explicitly obtained the physical wave - functions for @xmath0 , and have studied their features .
we considered the cases with and without a cosmological constant , and explicitly wrote down and solved the set of the partial differential equations representing the constraints . while the cases without a cosmological constant were more or less straightforward to solve in a standard method ( was indeed previously solved for @xmath7 in @xcite ) , the cases with a cosmological constant set a new challenge . in order to solve them
, we solved on an @xmath429-symmetric subspace in the configuration space for @xmath430 and then extended the solution over the whole space for @xmath7 . remarkably , the solutions in both the cases with and without a cosmological constant turned out to have the same form when expressed with some functions in which all the effects of the cosmological constant are included .
this actually hints that the inclusion of the cosmological constant does not change the structure of the system and tickles our intuition that this can be true even for higher-@xmath1 tensor models . in order to investigate this issue we went through a different path in the last sections .
since we did not assume any boundary ( or initial ) conditions , the solutions contain undetermined functions .
however , we were able to study the possible locations of the peaks of the wave - functions , which would be physically interpreted as enhanced configurations .
we have found that , for @xmath2 with no cosmological constant , the configurations invariant under an @xmath431 symmetry as well as those associated to a grassmann algebra ( this also appears for @xmath7 ) can be enhanced .
moreover , the intersecting points of these two cases have additionally a lorentz boost symmetry . while the @xmath431 and lorentz boost transformations commute with each other for @xmath2 , and thereby do not give rise to any interesting pictures , we expect more non - trivial phenomena to arise in higher-@xmath1 tensor models .
the second is that we have systematically constructed some series of solutions valid for any @xmath1 from the perspective of statistical systems on random networks , or random tensor networks .
it was noticed that the canonical tensor model shares some resemblances with the statistical systems on random networks , where the hamiltonian vector flow of the canonical tensor model corresponds to the renormalisation group flow of the statistical system @xcite .
the `` grand - type '' partition function of the statistical system satisfies the momentum constraints of the tensor model .
however , it does nt satisfy the hamiltonian constraints .
this was cured by integrating the `` grand - type '' partition function in a scale - free manner , which provided a physical wave - function of the canonical tensor model with no cosmological constant .
this is a remarkable achievement , as it gives a series of solutions valid for any @xmath1 .
an important fact noted in the analysis is that the solutions constructed in this manner are dynamical at least for @xmath432 .
an interesting connection between the `` grand - type '' partition function and an extension of airy function was also noticed by transforming the integration variables . the above process of constructing physical wave - functions was simplified and generalized .
we have constructed more general solutions valid for any @xmath1 and no cosmological constant in terms of simpler expressions with integration over an @xmath1-component vector .
these solutions are dynamical at least for @xmath433 as above .
a similar process of constructing wave - functions for no cosmological constant has been performed in the conjugate ( momentum ) representation . however , these solutions turned out to be kinematical .
then we also considered integration over matrix and tensor variables , and found a new dynamical solution as well as kinematical solutions .
we pointed out that , though kinematical solutions might be of less physical interest than dynamical ones , the former can extend the variations of the latter with continuous parameters .
the third is that we have performed a general analysis of the wave - functions in the momentum representation for the cases with a cosmological constant .
we started with the similar assumptions about the forms of wave - functions and solved the constraint equations .
remarkably , the solutions for @xmath9 with a cosmological constant turned out to be obtained by restricting the domains of the solutions for @xmath10 with no cosmological constant
. then we have proven a general theorem that the latter can always generate the former by such restriction .
this suggests that the effect of the cosmological constant can be absorbed into the dynamics of the tensor model with increased @xmath1 .
this is somewhat like in general relativity , where a de sitter background ( maybe arising due to a cosmological constant ) can be studied by embedding it in a flat space - time of one higher dimension .
this property is also true for more general solutions .
it is now evident that there exist at least some tractable solutions valid for any @xmath1 in the canonical tensor model , though the quantized constraints on physical states are non - linear and complicated .
it would be an interesting question how further we can proceed : is it possible to write down all the solutions in some tractable manners ? on the other hand , from the physical point of view , it would rather be important to obtain a unique physical state which satisfies a physically required initial condition .
then , one can discuss not only the possibilities but the real configurations which become the peaks of the wave - function .
this will inevitably be related with the space - time emergence in the canonical tensor model , and as explicitly shown for @xmath2 , there would be symmetry enhancement , which would be related with why the universe is a homogenous space - time with various gauge symmetries .
this enhancement would also suggest possible connections between the canonical tensor model and the group field theories@xcite , in which lie groups are embedded as input structures .
then , what is the initial condition ?
this question would also be tightly related to what is time in the canonical tensor model , which contains no preferred time variable in its basic formulation .
finally , the discussions on a cosmological constant in this paper strongly suggests that it can totally be ignored .
this would be a really interesting possibility in view of the so - called cosmological constant problem , and we hope to complete the discussions in future study .
gn would like to thanks ns , shinya aoki and others at the yukawa institute for theoretical physics ( yitp ) in kyoto university for providing the wonderful hospitality where a part of the work was done , and is grateful to them for providing the generous financial support .
ys is very grateful to yitp , where a part of this work was done , for their warm hospitality , and would like to acknowledge financial support from yitp : the bilateral international exchange program ( biep ) , and the long - term international exchange program for young researchers .
j. ambjorn , b. durhuus and t. jonsson , `` three - dimensional simplicial quantum gravity and generalized matrix models , '' mod .
lett . a * 6 * , 1133 ( 1991 ) . n. sasakura , `` tensor model for gravity and orientability of manifold , '' mod .
lett . a * 6
* , 2613 ( 1991 ) .
n. godfrey and m. gross , `` simplicial quantum gravity in more than two - dimensions , '' phys .
d * 43 * , 1749 ( 1991 ) .
p. di francesco , p. h. ginsparg and j. zinn - justin , `` 2-d gravity and random matrices , '' phys .
rept . * 254 * , 1 ( 1995 ) [ hep - th/9306153 ] .
r. de pietri and c. petronio , `` feynman diagrams of generalized matrix models and the associated manifolds in dimension 4 , '' j. math .
* 41 * , 6671 ( 2000 ) [ gr - qc/0004045 ] .
r. gurau , `` colored group field theory , '' commun .
phys . *
304 * , 69 ( 2011 ) [ arxiv:0907.2582 [ hep - th ] ] .
r. gurau and j. p. ryan , `` colored tensor models - a review , '' sigma * 8 * , 020 ( 2012 ) [ arxiv:1109.4812 [ hep - th ] ] .
v. bonzom , r. gurau , a. riello and v. rivasseau , `` critical behavior of colored tensor models in the large n limit , '' nucl .
b * 853 * , 174 ( 2011 ) [ arxiv:1105.3122 [ hep - th ] ] .
r. gurau and j. p. ryan , `` melons are branched polymers , '' arxiv:1302.4386 [ math - ph ] .
s. dartois , r. gurau and v. rivasseau , `` double scaling in tensor models with a quartic interaction , '' jhep * 1309 * , 088 ( 2013 ) [ arxiv:1307.5281 [ hep - th ] ] .
w. kamiski , d. oriti and j. p. ryan , `` towards a double - scaling limit for tensor models : probing sub - dominant orders , '' new j. phys .
* 16 * , 063048 ( 2014 ) [ arxiv:1304.6934 [ hep - th ] ] . v. bonzom and f. combes , `` fully packed loops on random surfaces and the 1/n expansion of tensor models , '' arxiv:1304.4152 [ hep - th ] .
r. gurau , `` the 1/n expansion of tensor models beyond perturbation theory , '' commun .
phys . * 330 * , 973 ( 2014 ) [ arxiv:1304.2666 [ math - ph ] ] .
s. dartois , v. rivasseau and a. tanasa , `` the @xmath5 expansion of multi - orientable random tensor models , '' annales henri poincare * 15 * , 965 ( 2014 ) [ arxiv:1301.1535 [ hep - th ] ] .
v. bonzom , `` new 1/n expansions in random tensor models , '' jhep * 1306 * , 062 ( 2013 ) [ arxiv:1211.1657 [ hep - th ] ] .
a. tanasa , `` multi - orientable group field theory , '' j. phys .
a * 45 * , 165401 ( 2012 ) [ arxiv:1109.0694 [ math.co ] ] .
s. dartois , `` a givental - like formula and bilinear identities for tensor models , '' arxiv:1409.5621 [ math - ph ] . v. a. nguyen , s. dartois and b. eynard , `` an analysis of the intermediate field theory of @xmath434 tensor model , '' arxiv:1409.5751 [ math - ph ] .
s. carrozza , `` discrete renormalization group for su(2 ) tensorial group field theory , '' arxiv:1407.4615 [ hep - th ]
. v. rivasseau , `` the tensor track , iii , '' fortsch .
* 62 * , 81 ( 2014 ) [ arxiv:1311.1461 [ hep - th ] ] .
s. carrozza , `` tensorial methods and renormalization in group field theories , '' arxiv:1310.3736 [ hep - th ] .
j. ben geloun , `` on the finite amplitudes for open graphs in abelian dynamical colored boulatov - ooguri models , '' j. phys .
a * 46 * , 402002 ( 2013 ) [ arxiv:1307.8299 [ hep - th ] ] .
j. ben geloun , `` renormalizable models in rank @xmath435 tensorial group field theory , '' arxiv:1306.1201 [ hep - th ] .
d. ousmane samary , `` beta functions of@xmath436 gauge invariant just renormalizable tensor models , '' phys
. rev .
d * 88 * , no .
10 , 105003 ( 2013 ) [ arxiv:1303.7256 [ hep - th ] ] .
s. carrozza , d. oriti and v. rivasseau , `` renormalization of a su(2 ) tensorial group field theory in three dimensions , '' commun .
phys . * 330 * , 581 ( 2014 ) [ arxiv:1303.6772 [ hep - th ] ] .
s. carrozza , d. oriti and v. rivasseau , `` renormalization of tensorial group field theories : abelian u(1 ) models in four dimensions , '' commun .
* 327 * , 603 ( 2014 ) [ arxiv:1207.6734 [ hep - th ] ] . j. ben geloun and e. r. livine , `` some classes of renormalizable tensor models , '' j. math .
* 54 * , 082303 ( 2013 ) [ arxiv:1207.0416 [ hep - th ] ] .
j. ben geloun and d. o. samary , `` 3d tensor field theory : renormalization and one - loop @xmath437-functions , '' annales henri poincare * 14 * , 1599 ( 2013 ) [ arxiv:1201.0176 [ hep - th ] ] . j. ben geloun and v. rivasseau , `` a renormalizable 4-dimensional tensor field theory , '' commun .
phys . * 318 * , 69 ( 2013 ) [ arxiv:1111.4997 [ hep - th ] ] .
d. v. boulatov , `` a model of three - dimensional lattice gravity , '' mod .
lett . a * 7 * , 1629 ( 1992 ) [ arxiv : hep - th/9202074 ] .
h. ooguri , `` topological lattice models in four - dimensions , '' mod .
a * 7 * , 2799 ( 1992 ) [ arxiv : hep - th/9205090 ] .
de pietri , l. freidel , k. krasnov and c. rovelli , `` barrett - crane model from a boulatov - ooguri field theory over a homogeneous space , '' nucl .
b * 574 * , 785 ( 2000 ) [ arxiv : hep - th/9907154 ] .
l. freidel , `` group field theory : an overview , '' int . j. theor
. phys .
* 44 * , 1769 ( 2005 ) [ hep - th/0505016 ] .
d. oriti , `` the microscopic dynamics of quantum space as a group field theory , '' arxiv:1110.5606 [ hep - th ] .
s. horata , h. s. egawa , n. tsuda and t. yukawa , `` phase structure of four - dimensional simplicial quantum gravity with a u(1 ) gauge field , '' prog .
* 106 * ( 2001 ) 1037 [ hep - lat/0004021 ] .
j. ambjorn , k. n. anagnostopoulos and j. jurkiewicz , `` abelian gauge fields coupled to simplicial quantum gravity , '' jhep * 9908 * ( 1999 ) 016 [ hep - lat/9907027 ] .
n. sasakura , `` canonical tensor models with local time , '' int .
j. mod .
a * 27 * ( 2012 ) 1250020 [ arxiv:1111.2790 [ hep - th ] ] . n. sasakura , `` uniqueness of canonical tensor model with local time , '' int .
j. mod .
a * 27 * ( 2012 ) 1250096 [ arxiv:1203.0421 [ hep - th ] ] . n. sasakura , `` a canonical rank - three tensor model with a scaling constraint , '' int .
j. mod .
a * 28 * ( 2013 ) 1 [ arxiv:1302.1656 [ hep - th ] ] .
d. oriti , `` group field theory as the 2nd quantization of loop quantum gravity , '' arxiv:1310.7786 [ gr - qc ] .
r. l. arnowitt , s. deser and c. w. misner , `` canonical variables for general relativity , '' phys .
rev . * 117 * , 1595 ( 1960 ) ; r. l. arnowitt , s. deser and c. w. misner , `` the dynamics of general relativity , '' arxiv : gr - qc/0405109 .
n. sasakura , `` quantum canonical tensor model and an exact wave function , '' int .
j. mod .
a * 28 * ( 2013 ) 1350111 [ arxiv:1305.6389 [ hep - th ] ] .
n. sasakura and y. sato , `` interpreting canonical tensor model in minisuperspace , '' phys .
b * 732 * , 32 ( 2014 ) [ arxiv:1401.2062 [ hep - th ] ] .
n. sasakura and y. sato , `` ising model on random networks and the canonical tensor model , '' ptep * 2014 * , no . 5 , 053b03 ( 2014 ) [ arxiv:1401.7806 [ hep - th ] ] . n. sasakura and y. sato , `` exact free energies of statistical systems on random networks , '' sigma * 10 * , 087 ( 2014 ) [ arxiv:1402.0740 [ hep - th ] ] .
b. s. dewitt , `` quantum theory of gravity .
1 . the canonical theory , '' phys . rev . * 160 * , 1113 ( 1967 ) .
n. sasakura , `` tensor models and 3-ary algebras , '' j. math .
* 52 * ( 2011 ) 103510 [ arxiv:1104.1463 [ hep - th ] ] . | tensor models , generalization of matrix models , are studied aiming for quantum gravity in dimensions larger than two . among them ,
the canonical tensor model is formulated as a totally constrained system with first - class constraints , the algebra of which resembles the dirac algebra of general relativity . when quantized , the physical states are defined to be vanished by the quantized constraints . in explicit representations ,
the constraint equations are a set of partial differential equations for the physical wave - functions , which do not seem straightforward to be solved due to their non - linear character . in this paper , after providing some explicit solutions for @xmath0 , we show that certain scale - free integration of partition functions of statistical systems on random networks ( or random tensor networks more generally ) provides a series of solutions for general @xmath1 .
then , by generalizing this form , we also obtain various solutions for general @xmath1 .
moreover , we show that the solutions for the cases with a cosmological constant can be obtained from those with no cosmological constant for increased @xmath1
. this would imply the interesting possibility that a cosmological constant can always be absorbed into the dynamics and is not an input parameter in the canonical tensor model .
we also observe the possibility of symmetry enhancement in @xmath2 , and comment on an extension of airy function related to the solutions . |
the concept of spontaneous breaking of a continuous global symmetry with the attendant goldstone phenomenon of massless scalar excitations has deeply influenced many developments in field theory and statistical physics .
the existence of massless modes can already be demonstrated in landau - ginzberg mean field analysis ( or tree level , in the field theory jargon ) .
however , it is not sufficient to stop at this level .
precisely because they are massless , the quantum / statistical fluctuations of the goldstone bosons may significantly alter the conclusions drawn from mean field theory.proofs of the goldstone theorem @xcite in quantum field theory have traditionally been used only to establish the existence of massless poles coupling to the currents of broken symmetry .
they have not usually focussed on the correlation functions .
most of the applications in quantum field theory have been based on the understanding at tree level .
notable exceptions are the soft pion theorems @xcite combining pcac with current algebra.these are mostly applied to extract the leading infrared behaviour .
a general analysis of these issues based essentially only on symmetry arguments was given some time ago by weinberg @xcite .
in that article he also proposed methods to go beyond the leading order results .
he also applied renormalization group methods to understand some universal features of these results . quite recently , some of weinberg s ideas have been extensively used in the context of the so called `` chiral perturbation theory''@xcite .
in contrast , in statistical physics , powerful techniques have been developed for handling the fluctuations in this context .
these are of importance for quantum field theory also.some examples are : i)the finite temperature chiral symmetry restoring transition in qcd , where goldstone phenomenon in three eucilidean dimensions is of relevance @xcite ; ii)qcd at temperatures below the chiral transition temperature as well as zero temperature qcd , where goldstone phenomenon in 4 euclidean dimensions is of relevance , and iii)the problems that are handled by chiral perturbation theory at present , i.e. effects of pion loops on various processes @xcite . in this paper
we point out various complimentary concepts and techniques for handling the effects due to quantum fluctuations of goldstone bosons .
we develop techniques for computing correlation functions away from the infrared regime.we show that there are certain universal features , valid even in four space - time dimensions.our approach provides an alternate procedure for chiral perturbation theory which is closer in its spirit to what has been proposed by weinberg on the basis of rg equations .
this aspect is elaborated upon in section 7 .
often the paradigm in quantum field theory is what is seen from a tree level analysis of the o(n ) linear and non - linear @xmath3-models .
the linear @xmath3-model is given by the action @xcite , @xmath5 @xmath6 transforms as the fundamental representation of o(n ) . if @xmath7 , the minimum of the classical potential is at @xmath8 .
therefore the vev @xmath9 is no longer zero in a tree level analysis.choosing the vacuum @xmath10 , among the equivalent vacua , we get , @xmath11 where @xmath12 .
therefore at tree level , we have @xmath13 massless pions and a massive @xmath3 with a mass @xmath14 . precisely because the goldstone bosons are massless , loop corrections
may drastically alter the above picture .
now it is well known that infrared divergences due to massless fluctuations drastically alter the conclusions of a semi - classical analysis , especially if the spacetime dimension d is less than four @xcite.already , way back in 1940 , holstein and primakoff @xcite argued that not only the transverse mode ( i.e. @xmath15 ) but also the longitudinal mode ( @xmath3 ) is soft in a quantum ferromagnet in three dimensions .
their arguments can be understood , as follows , in terms of the non - linear @xmath3 -model .
the non - linear @xmath3- model is relevant to the infrared dynamics of goldstone bosons @xcite .a simple heuristic argument is that if @xmath3 is massive , then for the study of the infrared dynamics of the pions , we may use an effective lagrangian obtained by integrating over the heavy @xmath3 degree of freedom .
also for leading infrared behaviour we may retain only the lowest ( two ) derivative terms in the effective action.the result is simply the heisenberg model , @xmath16 with the constraint , @xmath17 .
statistical physics in d space dimensions is simply the euclidean quantum field theory in @xmath18 space - time dimensions . in statistical physics
@xmath19 is proportional to the temperature , whereas in quantum field theory it is related to the pion decay constant , @xmath20(in what follows , we set @xmath21 ) .
if the spins are ordered in the direction of @xmath22 , then eliminating this field using the constraint , the action can be written entirely in terms of @xmath23 .
they describe the spin waves or the transverse fluctuations.at the tree level for the @xmath15 propagator we get , @xmath24 we may also consider the correlations for longitudinal fluctuations i.e. of the field @xmath25 ( : refers to any ordering prescription that makes products of fields at the same point meaningful ) .
this means , @xmath26 .
we may expect the leading infrared behaviour to come from the first term .
this would mean that so far as the leading infrared behaviour is concerned , @xcite , @xcite @xmath27 .
if we ignore self - interactions of @xmath28 , and use the above propagator for @xmath28 we get , for the @xmath3 propagator , @xmath29 this gives , atleast for @xmath30,in momentum space , @xmath31 ignoring the self interactions of @xmath15 s will later be seen to be justified because these are soft in the infrared region a consequence of the soft pion theorems.the result is that in @xmath32 the @xmath3propagator is @xmath33 , to be contrasted with the @xmath15-propagator.this means that @xmath3 also has long range correlations .
this appears to negate the naive deduction of the non - linear @xmath3 model obtained by integrating over the heavy @xmath3 field in the linear @xmath3 model @xcite .
one could dismiss these results as artifacts of a 1-loop calculation .
in fact the infrared divergences grow with the loop order .
therefore it is imperative to tackle this issue in a framework that transcends loop expansion.there is an exactly solvable model , the berlin - kac model @xcite which goes beyond loop - expansion and again gives precisely the same diverging longitudinal susceptibility and also a massless @xmath3 in @xmath30 .
again it is not clear whether this is an artifact of the model . in order to get a handle on the infrared dynamics of the @xmath3 field , pokrovsky and patashinky @xcite proposed a `` conserved modulus principle '' ( cmp).this principle holds that the fluctuations are such as to maintain the modulus of the n- component order parameter . in mathematical terms cmp
states that + @xmath34 now the expectation value of @xmath35 is given by the ornstein- zernike form : @xmath36 it is convenient to use susceptibility to describe the infrared behaviour.the longitudinal and the transverse susceptibility in presence of an external magnetic field @xmath37 are given by , @xmath38 using eqns .
[ pp1 ] and [ pp2 ] , we get , for @xmath30 , @xmath39 thus the longitudinal suceptibility also diverges .
patashinsky and pokrovski have offered a proof for cmp in @xcite .
they express the change in the thermodynamic potential u computed to quartic order in the fluctuations as + @xmath40\ ] ] and interpret this as implying cmp .
though to quartic order the correct form of this variation , which differs from the one given above ( and which can be verified using the ward identities discussed later ) , is @xmath41 + { 1\over 2 \chi_{l}}[{(\delta \vec\phi^2)^2\over 4 \bar \phi^2}]\ ] ] the fact that cmp is equivalent to the correct statement that the extreme ir behaviour is governed by the non - linear @xmath3-model indicates the possibility of a better proof of cmp .
+ the new era for understanding the role of fluctuations of goldstone bosons started with the renormalization group techniques .
brezin , wallace and wilson @xcite computed the equation of state for o(n ) @xmath3model using the @xmath42expansion.they found important differences from the corresponding calculation for the n=1 case ( i.e. the ising model).this could be traced to the infrared singularities due to the goldstone bosons that are present everywhere in the coexistence region for @xmath43 .
in particular it meant that the conclusions of the tree level analysis could not be right ( atleast for @xmath30 ) . at the critical point ( @xmath44 ) ,
@xmath3 and @xmath15 are on the same footing .
both their susceptibilities diverge with vanishing magnetic field as , @xmath45 governed by the critical exponent @xmath46 .
the question of interest is the behaviour for @xmath47 as @xmath48.the assumption that the singularities due to goldstone bosons exponentiate in the @xmath42expansion suggested that , @xmath49 holstein -primakoff @xcite , berlin - kac @xcite , and patashinski - pokrovski @xcite models have all made the same prediction .
such a result has also been proved rigorously for o(2 ) case using correlation inequalities in ref .
@xcite .
a deeper understanding can be obtained by considering the renormalization group flows shown in the figure below : at the critical point , @xmath44 , the infrared dynamics is governed by the o(n ) non - gaussian fixed point @xmath50 when @xmath1.everywhere on the critical surface the renormalization group ( rg ) trajectories flow from the the o(n ) gaussian fixed point @xmath51 ( which is infrared unstable ) to the o(n ) non - gaussian fixed point @xmath50 .
for @xmath52 we have the symmetric phase with massive ( and degenerate ) @xmath3 and @xmath53.the infrared behaviour is governed by the trivial fixed point @xmath54 which corresponds to zero correlation length.@xmath47 presents a new feature compared to n=1 case ( i.e.the ising model ) .
the theory has long range correlations for all @xmath55 as a consequence of the goldstone phenomenon.hence the infrared behaviour must be determined by a fixed point of infinite correlation length.what is this fixed point ?
@xmath56 , the deviation from the critical temperature in statistical mechanics , is to be identified with the square of the renormalized mass @xcite in the linear @xmath3-model .
in particular they are parameters with positive mass dimensionality . under rg flow
they necessarily scale.thus , once @xmath57 , the flow is towards @xmath58.this means that we are interested in the zero temperature fixed point .
an indication as to the nature of this fixed point came from a study of @xmath59expansion of o(n ) non linear @xmath3model by brezin and zinn - justin @xcite .
the o(n ) non - linear @xmath3- model in d=2 is renormalizable .
even though the renormalized perturbation is in terms of the spin wave modes , mermin - wagner theorem @xcite requires that the system is never in a phase with spontaneously broken o(n ) symmetry .
a hint of this is contained in the renormalized perturbation theory : the theory is asymptotically free.the coupling constant grows with the distance so that the long distance structure of the theory could be drastically different from the perturbation theory spectrum.indeed the spectrum is believed be that of massive and degenerate @xmath3 and @xmath15 s in the fundamental representation of o(n ) .
a consequence of the asymptotic freedom in @xmath60 is that in @xmath61 , the @xmath62-function @xmath63 has a non - trivial zero ( say at @xmath64 ) in addition to the trivial zero , @xmath65 , corresponding to the theory of ( n-1 ) free scalars.at @xmath64 the theory is scale invariant and corresponds to the critical theory with n massless scalars and with anomalous dimensions ( denoted by @xmath66 ) .it should be noted in fig .
1 that the fixed point @xmath50 of the linear @xmath67-model is identified with the fixed point @xmath66 of the non - linear model .
this follows essentially from a remark made in @xcite that at @xmath50 the correlation functions of the linear model are identical to those of the non - linear model where in the latter one includes correlation functions of composite operators also .
the coincidence of these two fixed points can also be established in the @xmath4-expansion scheme @xcite . for @xmath68
, the coupling constant grows with the distance , and as in @xmath60 , the theory is in the unbroken phase .
@xmath69 is of interest to us.in this case the infrared flow is towards @xmath65 .
the infrared behaviour is of o(n-1 ) gaussian theory ( this fixed point is denoted by @xmath70 in fig.1 . ) and the ultraviolet behaviour is of the critical theory with anomalous dimensions .
the fact that the infrared behaviour for @xmath69 is governed by the o(n-1 ) gaussian fixed point gives exhaustive information on the effects of quantum fluctuations of the goldstone bosons.it means that the pions decouple from each other at low momenta , and their propagators behave as the free propagators for small momenta.it also allows us to obtain the non - leading ir - behaviour as well as the effects of an external magnetic field.in particular we now understand why various heuristic considerations all gave the correct answer : it was the fortuitious behaviour of a free massless theory.anything more complicated would have been much more difficult to handle .
upto what dimension d is the analysis of @xmath61 calculation correct?in particular is it applicable to d=4?the case d=4 is somewhat different because of the absence of the non - gaussian fixed point @xmath50.at the critical temperature the infrared behaviour is believed to be given by free massless theory of n scalars as the beta function is positive for small couplings ; there is no evidence for a non - gaussian fixed point.this appears to exclude the validity of the @xmath59 analysis to this case.we need an alternate technique to handle this situation.this is provided , for example , by the @xmath4 expansion .
the @xmath4 expansion provides results which agree with the above analysis for all d.(the berlin - kac model @xcite is a variant of this technique . )
it yields the non - gaussian fixed point for @xmath1 and shows its absence for @xmath2.in either case it shows that the leading infrared behaviour anywhere in the broken phase is given by n-1 free pions.the only handicap of this technique is that it does not provide non - anomalous dimension for the non - gaussian fixed point @xmath50 in the leading n approximation.however even this is overcome once one takes higher order ( in @xmath4 ) corrections .
there are also indications that the technique gives qualitatively correct and quantitatively reasonable results for all @xmath71 .
the fact that even in d=4 the pion dynamics is governed by the o(n-1 ) gaussian fixed point @xmath70 in the infrared has important implications.in particular the propagator of @xmath3 ( to be interpreted as the o(n ) partner of the pions ) has an universal logarithmic singularity in the infrared region , @xmath72 where @xmath73 is the renormalisation scale .
this follows from the same arguments as for @xmath30 , in particular because @xmath27 in the infrared.this feature is present whether @xmath3-propagator has a pole at some @xmath74 or not .
the precise location of the pole is a non - universal feature depending on the microscopic dynamics .
it could have interesting implications for nuclear forces at low energies in the sigma-channel ( i.e. , spin j=0,isospin i=0 ) as will be discussed elsewhere . in sec . 3 and 4
we develop a new formalism for @xmath75 which will allow for explicit calculation of all correlation functions even away from the infrared regime.we will explicitly demonstrate all the properties mentioned above.our formalism gives the same results in the leading infrared limit whether one starts with the linear or the non - linear @xmath3- model . that the pions behave like free massless bosons in the infrared limit , in any @xmath76 , is such a general phenomenon that there should be an easier way of understanding it .
a careful analysis of the proof of the goldstone theorem along with current algebra and pcac provide one such non - perturbative argument.another is the treatment based on ward identities given in sec 5.low energy theorems based on pcac and current algebra do imply that pions decouple from each other and with other matter at low momenta , but they do not directly imply that the pion has a canonical dimension in the infrared.current algebra requires that the current coupling to the pion has canonical dimension @xmath77 .
however , the canonical dimension ( for the @xmath78 behaviour of the pion ) is @xmath79 .
this only means that the pcac relating the two needs a dimensionful parameter , but does not directly relate the two dimensions .
consider the time ordered correlation function of a current in a broken direction and the corresponding @xmath15 field : @xmath80 we get , @xmath81 \mid 0 > \not = 0\ ] ] since @xmath82 = \sigma(0 ) $ ] has a non - zero expectation value .
this means for the fourier transform of this correlation function , @xmath83 this requires that in the spectral representation of the correlation function there is necessarily a @xmath84 contribution implying a _ simple pole _ at zero mass.any other singularity , such as a branch point will not satisfy the above equation .
the fact that the @xmath3 channel also has an infrared singularity is equally general and should be understandable in a general way . in sec .
5 and 6 we provide two proofs which rely essentially on symmetry arguments .
the first uses ward identities , and the second , the schwinger- dyson equations .
it was argued in sec .
2 that the leading term in @xmath4 expansion retains all important effects of the goldstone boson fluctuations .
analytic calculations are possible in this limit and therefore we can analyse such effects in detail.there is an extensive literature on this technique @xcite.in this section we provide a new arsenal to this technique.we define a generalized generating functional of one - particle irreducible ( 1pi ) vertices and obtain an explicit expression for it in the large n limit .
this is made possible by involving an auxiliary field .
our method provides the easiest way of extracting all correlation functions without doing a separate calculation for each @xcite .
we present results for both linear and non - linear @xmath3-models .
@xmath87 $ ] , the generating functional of connected green s functions , is given by , @xmath88 } = \int { \cal d } { \vec \phi } e^ { i(s[{\vec \phi } ] + { \vec h}.{\vec \phi})}\ ] ] where , @xmath89 connected green s functions of @xmath90 are obtained from various functional derivatives of @xmath87 $ ] .
@xmath91 $ ] , the generating functional of 1pi vertices , is obtained from @xmath92 $ ] by a legendre transformation .
@xmath93 = w[\vec h ] - \vec h.\vec \phi\ ] ] where , @xmath94}{\delta h_i(x)}\ ] ] the vacuum expectation value ( vev ) of @xmath90 is obtained as the value at which @xmath85 is minimized.the second functional derivative about this minimum gives the inverse propagator .
further , various functional derivatives of @xmath85 about this minimum give the 1pi vertices .
the connected green s functions can be extracted by building _ only _ tree diagrams using these propagators and vertices . in the linear @xmath3-model ,
introduce an auxiliary field @xmath95 to make the exponent bilinear in @xmath96 : @xmath88 } = \int { \cal d } \lambda { \cal d } { \vec \phi } e^{i \int d^dx ( \frac{1}{2 } ( \partial_{\mu } { \vec \phi}(x))^{2 } -\frac{1}{2 } \lambda(x)({\vec \phi}^2(x)-c^2 ) + \frac{1}{4u } \lambda^{2}(x)+{\vec h}(x ) \cdot { \vec \phi}(x)}\ ] ] now an integration over each @xmath97 , gives a @xmath98 factor .
therefore ,
@xmath88 } = \int { \cal d } \lambda e^{i \int d^dx ( \frac{1}{4u } \lambda^{2}(x)+ \frac{c^2}{2 } \lambda ( x ) ) -\frac{n}{2 } tr ln \nabla -i \frac{1}{2 } h.\nabla^{-1}.h}\ ] ] where ,
@xmath99 on defining , @xmath100 = n w[{\vec h } ] \label{redefs}\ ] ] we get , @xmath101 } = \int { \cal d } \lambda e^{i n [ ( \int d^dx(\frac{1}{4u } \lambda^{2 } + \frac{1}{2}c^2 \lambda ) ) + \frac{i}{2}tr ln \nabla - \frac{1}{2 } h. \nabla^{-1}.h]}\ ] ] we are interested in the limit @xmath102 , @xmath103 holding @xmath104 fixed . with the variables redefined as in eqn .
[ redefs ] , it is clear that this limit is dominated by the saddle point .
the fluctuations about the saddle point give contributions which are non - leading in n. thus the leading contribution is , @xmath105 = \int d^dx(\frac{1}{4u } \lambda^{2}(x)+\frac{1}{2}c^2 \lambda(x ) ) \\
+ \frac{i}{2 } tr ln \nabla - \frac{1}{2 } h. \nabla^{-1}.h \label{ggf2}\ ] ] where @xmath106 , the saddle point , is a functional of @xmath107 as given by , @xmath108 we may obtain @xmath91 $ ] from @xmath109 $ ] by a legendre transform as described earlier .
but this form is not easily amenable to further analysis .
instead we note that the saddle point equation may be rewritten as @xmath110 where , @xmath111 notice that these equations can be obtained from the following functional of @xmath112 and @xmath95 : @xmath113 & = & \int d^dx(\frac{1}{4u } \lambda^{2}(x ) + \frac{1}{2}c^2 \lambda(x ) + \frac{1}{2}(\partial_{\mu } \vec \phi)^2 ( x ) \nonumber \\ & & - \frac{1}{2 } \lambda(x ) \vec \phi^2(x ) ) + \frac{i}{2}tr ln \nabla \label{gammainf}\end{aligned}\ ] ] by taking functional derivatives : @xmath114 } { \delta \phi_i(x ) } , \quad - \kappa(x)=\frac{\delta \gamma_{\infty}[\vec \phi,\lambda ] } { \delta \lambda(x ) } \label{ggf1}\ ] ] here we have temporarily added a source @xmath115 for @xmath95.when restricted to correlation functions with only @xmath116 on external legs , @xmath115 can be set to zero , yielding eqn 30 .
these equations imply that @xmath117 $ ] and @xmath118 $ ] are related by a functional legendre transformation in both arguments .
this means that the connected green s functions of @xmath112 and in fact also of @xmath119 ( i.e. functional derivatives of @xmath117 $ ] wrt both @xmath120 and @xmath115 ) can be obtained from tree diagrams built from the action @xmath121$].this connection is simply an algebraic consequence of the legendre transformation .
we have thus the option of constructing green s functions involving the composite operator @xmath122 also by having @xmath95 as external legs .
by introducing an auxiliary field @xmath95 we have obtained an explicit form of the generating functional of 1pi vertices which is far more amenable to analysis than if we had obtained a @xmath123 by substituting the saddle point solution for @xmath95 .moreover we find that the properties of the phase with spontaneously broken symmetry can be understood in a simple way in terms of this auxiliary field .
ofcourse , now @xmath95 also appears in propagators and vertices .
the usual generating function @xmath124 $ ] can however be recovered by using @xmath125 from eq.[ggf1 ] to replace @xmath95 as a functional of @xmath126 in @xmath127 $ ] .
there is hardly any change in our calculations ( and results ) when passing over to the non - linear @xmath3-model .
now we have , @xmath128 } = \int { \cal d } \vec \phi \pi_x \delta^d(\vec \phi^2(x)-c^2 ) e^{i \int d^dx ( \frac{1}{2 } ( \partial_{\mu } \vec \phi(x)^{2 } + \vec h(x ) \cdot \vec \phi ( x)})\ ] ] using the fourier representation for the functional @xmath46 function , @xmath129 } = \int { \cal d } \lambda { \cal d } \vec \phi e^
{ i\int d^dx ( \frac{1}{2 } ( \partial_{\mu } \vec \phi(x)^{2 } ) -\frac{1}{2 } \lambda ( \vec \phi^2 ( x)-c^2 ) + \vec h(x ) \cdot \vec \phi(x))}\ ] ] in this form the only difference with the linear @xmath3-model is that a quadratic term in @xmath95 is missing in the exponent.equivalently , the non - linear @xmath3-model is simply the special case @xmath130 of the linear @xmath3-model .
thus all results for the non - linear @xmath3-model can be extracted by taking this limit in our formulae.we find that it is straightforward to take this limit and the results are not very different .
we have been cavalier in our manipulations with the functional integrals.for instance , the @xmath131 term has to be carefully defined.we may choose a pauli - villars or a lattice cut - off for defining this term.if we consider a functional taylor expansion of the tr ln term about a constant value of @xmath95 , for @xmath1 , only the term linear in @xmath95 is sensitive to the cut-off.we may therefore obtain a renormalized @xmath132 $ ] by making a specific cut - off independent choice for this term . in @xmath2 , in addition , the term quadratic in @xmath95 is also divergent and requires a renormalization.for the sake of uniformity , we shall introduce local counterterms @xmath133 for both @xmath32 and @xmath2 . in the @xmath32 case , this amounts to a finite renormalisation of u. because of the presence of massless modes , renormalisation should be so performed that it does not introduce spurious infra - red divergences .
hence we choose to define renormalised quantities at @xmath134(say ) .
then the renormalised effective action is @xmath113 & = & \int d^dx(\frac{1}{4u } ( \lambda(x)-\mu)^{2 } + \frac{1}{2}c^2 ( \lambda(x)-\mu ) + \frac{1}{2}(\partial_{\mu } \vec \phi)^2 ( x ) \nonumber \\ & & - \frac{1}{2 } \lambda(x ) \vec \phi^2(x ) ) + \frac{i}{2}tr^\prime_{\mu } ln \nabla \label{gammaren}\end{aligned}\ ] ] @xmath135 means that the * local * part of the first two terms in the functional expansion around @xmath134 are to be ignored.now u and c are renormalised quantities .
thus the model is renormalizable in the usual sense.it is also seen that the wave function renormalization for @xmath95 , equivalently the composite operator @xmath119 , is finite in the @xmath102 limit.(for a discussion of renormalizability of the @xmath4 series see ref .
@xcite ) .
the discussion of renormalizability is equally valid for the the non - linear @xmath3-model .
the only difference is that the bare @xmath136 term is not present in the action.the quantum effects introduce such a term which is finite in @xmath1 and diverges logarithmically with the cut - off in @xmath2.thus the non - linear @xmath3-model is not renormalizable in this sense in @xmath2 even in the @xmath137 limit .
in this section , we extract the infrared behaviour of the correlation functions , in particular , in the phase with sbs .
first we need to obtain the vev s @xmath138 and @xmath139 of the fields @xmath116 and @xmath95.for uniform ( in particular zero ) fields @xmath140 ( and @xmath115 ) , the vev s are space- time independent.therefore we get @xmath141 @xmath142 \label{mini}\ ] ] the analysis below closely follows ref .
we first consider the situation in the absence of the external magnetic field , @xmath143.now @xmath144 and so there are three cases : 1 .
@xmath145 and @xmath146 corresponds to the critical point.this is reached for the critical value @xmath147 where , + @xmath148 2 .
@xmath149 and @xmath150 is the symmetric phase .
@xmath139 is given by , + @xmath151 + where we have used eqn.[casei ] 3 .
@xmath152 , @xmath153 is the phase with sbs.the vev @xmath154 is given by , + @xmath155 + as seen by using eqn.[casei ] for any @xmath156 we have a solution for a real @xmath154 and this corresponds to a phase with sbs .
on the other hand for any @xmath157 we have a solution for @xmath139 in eqn.[caseii ] , corresponding to the symmetric phase . in terms of the fluctuations @xmath158
we have for the symmetric phase , @xmath113 & = & \int d^dx ( \frac{1}{2 } ( \partial_{\mu } \vec \phi(x)^{2 } ) -\frac{1}{2}\lambda_0 \vec \phi^2 ( x ) -\frac{1}{2}\lambda'(x ) \vec \phi^2 ( x ) ) \nonumber \\ & & + \frac{1}{4 } \int \int d^dx d^dy \lambda'(x)i(x - y,\lambda_0 ) \lambda'(y ) \nonumber \\ & & + \frac{i}{2}tr'_\mu ln ( -\partial^2- \lambda_0-\lambda')\end{aligned}\ ] ] where @xmath159 is the fourier transform of @xmath160+\frac{1}{u}\ ] ] all terms linear in the fluctuations drop out because we are expanding about the extremum value.@xmath139 is seen to be the @xmath161 of the @xmath126 field . for the phase with sbs we choose @xmath162 . in terms of the fluctuations ,
@xmath163 we get , @xmath164 & = & \int d^dx ( \frac{1}{2 } ( \partial_{\mu } \vec \phi(x))^{2 } -v \lambda(x)\sigma(x ) -\frac{1}{2 } \lambda(x ) ( \sigma^{2}(x)+\vec \pi^2(x ) ) \nonumber \\ & & + \frac{1}{4u } \int \int d^dx d^dy \lambda ( x)i(x - y,0 ) \lambda ( y ) + i \frac{1}{2}tr'_\mu ln ( -\partial^2-\lambda)\nonumber \\\end{aligned}\ ] ] @xmath165 is massless as is to be expected . the mixing of @xmath3 with the auxilary field @xmath95 is a crucial feature of this phase .
this mixing governs the infrared behaviour of the correlation functions as seen below .
the terms quadratic in @xmath3 and @xmath95 may be written in momentum space as , @xmath166 now , @xmath167 we get the propagators by inverting the matrix of the quadratic form in eqn.[quad ] .
@xmath168 we are now in a position to extract the infrared behavior of various correlation functions .
the propagators have the following infrared behaviour : @xmath169 thus the leading n approximation reproduces the correct infrared behaviour for the @xmath15 and the @xmath3 propagators , viz . , @xmath15 has canonical @xmath170 behaviour , whereas @xmath3 has an anomalous dimension @xmath0 , precisely as if @xmath171.we now demonstrate explicitly that the other correlation functions also have the infrared behaviour required by the soft pion theorems .
the interactions of @xmath15 and @xmath3 are only mediated through the auxiliary field @xmath95 in this formalism , and the @xmath95 propagator has an unusual @xmath172 infrared behaviour.this is the essential reason why @xmath15 s and @xmath3 s decouple from each other in the infrared , and the results of soft pion theorems are valid . for the leading ir - behaviour ,
the multi-@xmath95 vertices are irrelevant .
the effective action is then quadratic in @xmath95 and consequently , the @xmath95-field can be exactly integrated out.this leads to the following effective action to describe the leading ir behaviour : @xmath173\ ] ] this summarises all the essential results of this paper .
thus the large n approximation provides a valuable non perturbative tool for obtaining effects due to quantum fluctuations of goldstone bosons.it incorporates the infrared effects correctly.it also incorporates the ultraviolet behavior correctly : for @xmath1 , the ultraviolet behaviour is of the non - gaussian fixed point , and for d=4 , it is of the o(n ) gaussian fixed point .
it has an explicit @xmath3 field , so that effects due to its infrared behaviour on other matter can be investigated .
also the technique provides a way of going beyond the leading infrared region.it provides a detailed structure for the propagators of both @xmath15 and @xmath3 with correct asymptotic behaviours .
we now consider the situation with a uniform external magnetic field @xmath174 .
now , @xmath175 implies that neither @xmath139 nor @xmath176 is zero.this is the choice that removes terms linear in the shifted field @xmath3 one gets essentially the same @xmath177 as the one we got for the symmetric phase but with a crucial additional term @xmath178 and with the proviso that elsewhere @xmath119 now stands for @xmath179 .
thus @xmath15 has a mass @xmath180 .
@xmath3 has a mixing with @xmath95 as in the sbs case.the @xmath181 propagators may be computed as before .
the equation governing @xmath139 is now , @xmath182 with @xmath183 . by an analysis of these equations one obtains a mass that scales as @xmath184 for the pions and as @xmath185 for the sigma . for applications to statistical mechanics
, one should use the expressions derived here continued to the relevant number of euclidean dimensions .
in this section we demonstrate that the @xmath186 diverges for @xmath187 , that @xmath188 in the infrared regime and the softness of multipion scattering amplitudes using ward identities .
chiral ward identities merely reflect the underlying symmetry structure of the theory and are useful both in the symmetric phase where they relate correlation functions of the same type ( n - point functions with fixed n ) as well as in the ordered phase where they relate different n - point functions in a non - trivial way .
note that in the generating function both the measure @xmath189 , and the action @xmath190 , are invariant under @xmath67 rotations generated by @xmath191;the source term is also invariant under a simultaneous rotation of both the source @xmath192 and the @xmath112 field.considering an infinitesimal transformation , this implies for @xmath85 , @xmath193 = 0 \label{b}\ ] ] when an external field @xmath192 is present.this means that @xmath85 has only o(n ) invariant combinations such as @xmath194 etc .
, except for one additional term @xmath195 , depending on the external field . by successive differentiations of this identity w.r.t .
the fields @xmath196 , one can generate various forms of ward identities.in particular , an identity that will prove most useful later on is @xmath197 where @xmath198 is the expectation value of the order parameter .
+ to explore the infrared properties , we consider the thermodynamic potential , which is called the the exact effective potential in quantum field theory .
we include an external magnetic field to serve as an infrared regulator and consider the singular behaviour as this field goes to zero .
the various zero momentum n - point functions are obtained by functionally differentiating this potential w.r.t .
constant fields and evaluating the derivatives at the minima of the potential . in the ordered phase
this minima occurs at a non - vanishing value @xmath199 as the magnetic field is reduced to zero .
+ as a consequence of the ward identity , eqn.[b ] , the effective potential has the form , @xmath200 since @xmath201 the vev @xmath154 is given by , @xmath202 where h is the magnetic field along the @xmath3-direction .
we also have @xmath203 @xmath204 @xmath205 etc . here
@xmath206 etc .
are the derivatives of @xmath207 w.r.t . its argument .
since @xmath208 we get , @xmath209 also , @xmath210 it can be seen that eqns .
[ d ] , [ c ] and [ e ] together are equivalent to the @xmath78 limit of the identity in eq.[a ] .
> from eq.[f],the 1pi pion four - point vertex at zero momentum is given by @xmath211 and from eq.[e ] , the @xmath212 1pi vertex is given by @xmath213 hence the @xmath214-scattering amplitude at zero momentum @xcite is , @xmath215\ ] ] in other words @xmath216 irrespective of @xmath217 , @xmath218 is at least as soft as h. this is equivalent to the soft pion theorems of current algebra @xcite .
note that this result has been proven here without the detailed assumptions of current algebra and pcac .
+ on introducing the shorthand notation @xmath219\ ] ] and @xmath220 one can write the general result @xmath221 now let us consider the zero momentum 1pi vertex function @xmath222 ( with @xmath223 ) corresponding to n - longitudinal and m - transverse external legs.then all tensors @xmath224 will vanish when @xmath225 as then at least some of the @xmath126 s in these tensors will have to be transverse.furthermore the leading ir behaviour is equivalent to treating @xmath154 as being very large ( compared to the momenta and equivalently to h ) .
it is clear that due to the exchange of the massless goldstone particles in the intermediate state , the n - point functions should have infrared singularities for sufficiently large @xmath226 . from very general considerations , it follows that the nature of these singularities becomes more severe with increasing @xmath226 .
thus , for sufficiently large @xmath227 , the derivative @xmath228 has the structure @xmath229 where @xmath230 therefore @xmath231 now three situations can arise : ( a ) @xmath232 is finite or goes to zero slower than @xmath233 as @xmath234 .
then @xmath235 is more singular than @xmath236 .
also @xmath237 from ( 67 ) : ( b)@xmath232 and @xmath233 have the same leading infrared behaviour .
further , ( 63 ) already implies that @xmath238 and @xmath239 which is a rather strong result .
then all @xmath240 for @xmath241 have the same infrared behaviour .
we shall show shortly that this is inconsistent .
( c ) @xmath242 vanishes faster than @xmath233 as @xmath243 .
this has the consequence that @xmath244 has softer ir singularity than @xmath245 for @xmath246 . by virtue of ( 63 ) this choice also implies that @xmath247 .
this choice also implies that @xmath218 vanishes faster than @xmath248 from ( 67 ) . as these features are already invalidated in perturbation theory , ( c )
will not be considered any more .
now , for both the situations a ) and b ) it follows that the dominant contribution to @xmath222 will come from @xmath249 in ( 70 ) ; this is the maximum possible value for @xmath250 .
hence @xmath251 this further implies that @xmath252 this means that as far as the infrared behaviour of all the vertex functions is concerned @xmath253 this and the fact that pions are non - interacting in the ir leads to the stated result on @xmath186 _ viz .
_ @xmath254 for @xmath1 and @xmath255 for @xmath2 .
but these behaviours for @xmath186 are inconsistent with the situation ( b ) which demand @xmath256 .
thus only ( a ) need be considered . in order to get an appreciation for the nature of the effective potential , we analyse it in the large n limit.from the generating functional @xmath257 $ ] of sec.3 .
one can recover the usual generating functional @xmath258 $ ] by eliminating @xmath95 using its equation of motion . since we are interested in constant fields @xmath259 , it is sufficient to have the @xmath95 field also constant .
subtracting eqn . [ mini ] from the analog of eqn .
[ saddle ] that results from the renormalised @xmath177(eqn [ gammaren ] ) , we get , @xmath260 = ( \vec\phi^2 - v^2)\ ] ] which can be solved to yield the implicit solution @xmath261 for given @xmath262.therefore the effective potential has the form , @xmath263 where @xmath95 is to be substituted as in eqn .
further we know that the minimum of the effective potential is at , @xmath264 such that @xmath265 , as @xmath266.this means that @xmath95 and also @xmath267 when expanded about the minimum @xmath268 has taylor coefficients which are increasingly singular in @xmath139 and hence in @xmath269 .
in this section we present an argument based on schwinger - dyson equations that shows the diverging @xmath186 as a consequence of the infrared divergence of the goldstone bosons .
the heirarchy of schwinger - dyson equations are obtained by exploiting the following property of functional integrals @xmath270 to be specific let us consider @xmath271\ ] ] the above identity implies the following equation for @xmath85 , the generating functional of 1pi diagrams : @xmath272\end{aligned}\ ] ] where @xmath273 and @xmath274 it should be noted that @xmath275 and @xmath276 are the propagator and 3-point vertex respectively only when the above derivatives are evaluated at the minimum of @xmath85 namely , at @xmath277 such that @xmath278 hence @xmath279 and @xmath280 in eqs([dinv ] ) and ( [ vinv ] ) are @xmath126-dependent .
+ it will prove useful to display explicitly the behaviour of the 1pi graphs in the large
for example , the four - point vertex has a @xmath4 dependence for large n. likewise , the expectation value of the order parameter has a @xmath281 behaviour , and hence the 3-point functions in the ordered phase will have a @xmath282 factor . now consider the sd equations for @xmath283 .
this equation is best represented diagrammatically as in fig .
2 . in these figures
solid lines represent exact @xmath3-propagators and dashed lines exact @xmath15-propagators ( except the lines with subscript ` 0 ' denoting tree level propagators ) , and all blobs represent exact vertices . + we assume some infrared regulator like the external field h. then @xmath284 , the goldstone boson mass is not zero and we can safely study the zero momentum behaviour of various graphs .
the degree of infrared divergence is reflected by the leading @xmath284-dependence .
our input is that the exact pion propagator has the behaviour @xmath285 for @xmath78 , i.e. we put in the information that the pion has a canonical behaviour in the infrared and study the consequences for the @xmath3-propagator.diagram 2(a ) contributing to @xmath283 would behave like @xmath286 if @xmath287 were o(1 ) in the ir ( and consequently the total @xmath283 would also be o(1 ) due to the ward identity ( eqn [ a ] ) ) .
this contribution is of order o(1 ) in the large - n count .
thus , either the ir - singular behaviour of 2(a ) is cancelled by other diagrams in fig ( 2 ) or * @xmath287 must be at least as soft as @xmath288*. diagrams with potential ir - divergences are 2(b),2(c ) and 2(d ) ; apart from all of them being of o(1/n ) , none of them is as ir - singular as 2(a ) .
hence cancellation is ruled out and @xmath287 must behave at least like @xmath289 with @xmath290-positive semidefinite . consequently @xmath283 must vanish as either @xmath291 or @xmath289 whichever is dominanat . + while this proves conclusively that @xmath186 must diverge , we still need to restrict @xmath290 further .
it should be noted that @xmath283 has terms of order @xmath292 and @xmath293 in the large - n count . on
the other hand @xmath283 must vanish as some power of @xmath284 .
this is only possible if the @xmath292 terms are cancelled by such terms from the other diagrams .
as all the diagrams except 2(a ) are @xmath294 , generically this can only happen from the ultraviolet contributions to 2(a ) .
within @xmath4 perturbation theory at least , such cancellations are not possible in which case the @xmath292 terms must be generated in the infrared by 2(a ) and this is possible only if @xmath295 .
this completes the proof based on sd - equations .
we have thus established that the divergence of the longitudinal susceptibility is very general depending only on the goldstone phenomena .
in fact , we have shown on very general grounds that the leading infrared behaviour of the longitudinal two - point function is completely universal .
now we show that the same considerations lead to certain universal features of multipion scattering amplitudes .
such universal features were alluded to in weinberg s article @xcite .
consider , for example , the expression ( 67 ) for the @xmath214 scattering amplitude at zero momentum .
eqn ( 74 ) can be recast as @xmath296 hence @xmath297 as @xmath298 and one has @xmath299 on recalling that in the absence of the infrared cut - off h , the scale that replaces @xmath248 is @xmath300 , we recognise that the leading universal infrared behaviour @xmath301 is indeed the well known result for soft pion scattering ( usually derived in @xmath2 )
. this universal behaviour does not depend on the space time dimensionality . however ( [ new])demonstrates that even the next to leading term is universal .
the above considerations can be applied to multi - pion scatering amplitudes and one again finds the same universality of the first two terms .
it is also clear from this paper that these results are non - perturbative in nature .
the presence of these non - analytic ( in energy - momentum variables ) terms in @xmath2 had been noted by li and pagels @xcite , the coefficients of which in the case of @xmath214 scattering had been calculated by lehmann , and lehmann and trute @xcite using unitarity .
weinberg showed that the application of renormalisation group can be used to show the universality of these logarithmic terms .
he implicitly uses chiral symmetry .
though superficially it appears that the universality of the logarithmic terms is a one - loop result , careful reexamination of weinberg s demonstartion shows that it is an all - loop result . in our approach , the origin of the universality of these subdominant logarithmic terms
is directly related to universal divergence of the longitudinal susceptibility .
our large - n derivation , which links the divergence of the @xmath186 to the special properties of the auxiliary field also demonstrates the universal features in a very straightforward manner .
the multipion scattering amplitudes in this case can be classified into three categories : a ) those involving only @xmath95 and @xmath15 propagators , and the @xmath302 vertices . for the scattering amplitude involving 2n pions ( including initial and final states )
, one needs @xmath303 and @xmath304 propagators .
hence the ir - behaviour of all multipion amplitudes of this class is @xmath305 .
b ) amplitudes involving one @xmath306 vertex and n @xmath302 vertices .
the @xmath306-vertex scales like @xmath307 .
hence this contribution to the amplitude scales like @xmath308 , which is less dominant than the behaviour a ) .
finally , there is the class c ) which consists of amplitudes involving combinations of multi-@xmath95 vertices with @xmath309.their contribution can be seen to be even less dominant than b ) .
because of the @xmath310 factors in the denominators of the @xmath95-propagator , one gets a universal correction factor @xmath311 $ ] and the leading ir behaviour of these correction factors is the same in the linear and non - linear @xmath3-models .
this is precisely the same quantity that governs the universal divergence of @xmath186 .
we now make a few geenral remarks about conventional chiral perturbation theory ( chpt ) and compare it with our approach .
as we have already stated many times earlier , the leading ir behaviour of systems with goldstone bosons is completely universal , and in the sector of goldstone bosons only , is represented by the lagrangian @xmath312 this , following weinberg , can be cast in the form @xmath313 the validity of ( [ new-1 ] ) is , however , strictly in the extreme infrared .
one of the procedures adopted to extend ( [ new-1 ] to regions beyond the extreme ir is the so - called chiral perturbation theory . in the form that is mostly practised ,
chpt amounts to starting , for example , from a lagrangian of the form @xmath314 contrary to the attitude of the early days of effective lagrangians , one now carries out loop calculations with lagrangians of the type ( [ new-2 ] ) and ( [ new-3 ] ) . owing to perturbatie non - renormalisability of these lagrangians
, one needs infinitely many counter terms to absorb the various resulting infinities .. however , to any any desired accuracy in the momenta involved , one needs only a finite number of counter terms .
the structure of the resulting renormalised lagrangian whose domain of validity is somewhat larger that that of ( [ new-1 ] ) is that it is charaterised by a number of unknown parameters corresponding to the number of counteer terms used as well as certain universal parts independent of these unknown coefficients ( @xmath315 is always treated as known ) .
this is indeed the general structure argued by weinberg .
our analysis that the leading ir behaviour is governed by the gaussian fixed point @xmath70 points to a more efficient way of realising the results of chpt .
this suggestion was implicit in @xcite also .
this consists of perturbing the ir fixed point directly with the help of irrelevant operators .
their influence on all correlation functions can be directly computed by the use of the rg techniques .
the ultraviolet degrees of freedom have no role to play in this scheme .
this method will be applied to practical calculations of chpt elsewhere .
in this paper we have emphasised that the quantum fluctuations of the goldstone bosons have some drastic effects which are not encountered either in a tree level or loopwise analysis of the linear or the non - linear @xmath3- models.nevertheless,the leading infrared behaviour of the correlation functions again become identical in the linear and the non - linear versions.it is remarkable that the longitudinal fluctuations also become ` massless ' as a consequence of the masslessness of the goldstone bosons .
we have exhibited the generality of this phenomenon giving three separate arguments based on i ) .
renormalization group flows , ii ) .
ward identities , iii ) .
schwinger - dyson equations .
heuristic derivations of the soft - pion theorems within the linear @xmath3-model are based on the fact that @xmath3-mass ( @xmath316 ) being non zero , for all processes involving momenta much smaller than this mass the @xmath3-propagator can be replaced by @xmath317 .
one may wonder whether the results of this paper would vitiate this and consequently the soft - pion theorems too .
but as demonstrated in the sections on the large - n analysis as well as the one on ward identities , the soft pion theorems are still valid .
the deeper reason for this is , of course , the renormalisation group point of view according to which the leading ir - behaviour is controlled by the gaussian fixed point . +
as stressed sufficiently in this article , both the qualitative and quantitative aspects of the result , as far as the susceptibilities are concerned , were known in the literature from somewhat differing perspectives which were not general enough .
what we have done here is to demonstrate the full generality of this result as well as extend the analysis to include the nonperturbative ir behaviour of correlation functions also . despite the fact that the result concerning @xmath186 had appeared in the literature in different guises
, its significance does not appear to have been well appreciated .
+ we emphasise that in the chiral limit the ` sigma particle ' has a singular infrared behaviour ( i.e. logarithmic divergence ) even in four dimensions . independently of the space - time dimensions ,
the goldstone bosons have canonical dimensions and decouple from each other and their logitudinal partners in the infrared.we pointed out that this can be understood from the goldstone theorem , current algebra and pcac on the one hand and the flow towards @xmath318 gaussian fixed point on the other hand .
this latter point of view is very valuable because one can easily study the non - leading corrections to the infrared behaviour by standard renormalization group methods @xcite .
all effects of the microscopic theory could be parametrised in a few irrelevant and marginal operators around the @xmath318 gaussian fixed point and their effects computed .
this approach provides a technically more efficient and conceptually simpler alternative to chiral perturbation theory , which focusses on the infrared divergence of the goldstone bosons instead of the non - renormalizable ultraviolet divergences .
we showed that the leading @xmath319 approximation takes into account all the infrared behaviours accuratly , including that of the longitudinal fluctuations , in contrast to perturbation theory based on the linear or the non - linear @xmath3- models.we developed a technique which allows explicit computation of the correlation functions even far away from the infrared region.this provides a useful non - perturbative technique for obtaining a qualitative understanding of the effects of fluctuations .
a convenient tool for understanding longitudinal susceptibility is the equation of state @xcite.in terms of the reduced variables @xmath320,where m is the magnetisation , t the deviation from the critical temperature , and @xmath321 the standard exponents , the equation of state takes the universal form + @xmath322 as h tends to zero , @xmath323 must tend to one of the roots of @xmath324 . with suitable normalisation this root ( corresponding to spontaneous magnetization ) can be arranged to be at @xmath325.now , the question of whether @xmath186 diverges or not can be answered by knowing the manner in which @xmath324 approaches zero as @xmath323 approaches @xmath326.if this approach is parametrised as + @xmath327 where p is a positive semidefinite number , the behaviour of @xmath186 as h tends to zero is given by + @xmath328 in the large - n expansion for arbitrary d ( except d=4 where there are logarthmic modifications),the equation of state is @xmath329 as can be proven using the euclidean version of eqn([state ] ) of sec 4(see also @xcite ) . as the exponent of @xmath330 is greater than 1 one concludes that the longitudinal susceptibility diverges for all @xmath331 .
modulo the wavefunction renormalisation constant , @xmath186 is just @xmath332 .
there are interesting consequences of this phenomena both for finite temperature qcd as well as qcd at zero temperature . in the case of the former , under the well argued scenario that this is a second order transition , wilczek and rajagopal @xcite conclude that the transition must belong to the universality class of @xmath333 magnet models .
then the general results of this paper become applicable .
+ in numerical simulations of the qcd chiral phase transition at finite temperature , the quantity that has been used to characterise the nature of the phase transition is the so - called @xmath334-cumulant defined by @xcite @xmath335 where @xmath336 is the chiral condensate and @xmath337 the current - quark mass.thus @xmath338 below @xmath339,but close to it , one should expect , on the basis of the considerations of this paper @xmath340 the same considerations apply to the spontaneously broken phase of zero temperature qcd . here
one should find @xmath341 characterstic of goldstone phenomena in four dimensions .
+ of course , the arguments of wilczek and rajagopal that maps the problem to a classical stat mech problem in @xmath32 is valid exactly at @xmath339 . but continuity would demand that at temperatures in the vicinity of @xmath339 too the effective dimensionality should be close to 3 .
thus , the expectations based on eqs ( [ cum1],[cum2 ] and [ cum3 ] ) are the correct ones .
this should yield the experimentally interesting signal that for qcd at finite temperatures @xmath331 , @xmath186 should smoothly go over from eqn ( [ cum2 ] ) to eqn([cum3 ] ) as t is lowered .
+ indeed , wherever there are goldstone bosons , one should expect to see the singular behaviour of @xmath186 .
it is therefore somewhat disappointing that clear experimental signatures of this effect have not yet been found even for the classic ferromagnets.the lack of isotropy in most real life ferromagnets would make the experimental establishment of this effect a challenging one .
an attempt to see this for ferromagnets has recently been made by @xcite .
they should have been looking for a @xmath342 behaviour rather than @xmath343 .
the fact that the behaviour of @xmath186 is super - universal in the sense that it depends only on dimension and the existence of goldstone modes should be established in as many diverse experimental circumstances as possible . as patashinsky and
pokrovsky @xcite have pointed out , the effect should also be observable in the superfluid phase of he.numerical simulations of the @xmath67- models in three dimensions , performed in the ordered phase , should also establish this effect .
+ kogut _ et al . _
@xcite report a completely different behaviour and different exponents for the finite temperature chiral transition in qcd in comparison to @xcite .
in particular , kogut _ et al . _ do not find the crucial softening of the longitudinal mode .
more specifically , they argue that pcac restricts the equation of state to be of the form + @xmath344 there is a fallacy in their argument ; pcac implies the above form of the equation of state if and only if the longitudinal susceptibility is assumed to be finite .
only then a taylor expansion around zero applied magnetic field is permissible.but our results about the divergence of @xmath186 imply that the assumption of a finite @xmath186 is incorrect .
once allowance is made for a diverging @xmath186 , pcac no longer places any restriction on the form of the equation of state . also , kogut _ et al . _
carry out their analysis in @xmath2 even at the critical temperature .
+ * acknowledgements * + the main inspiration for this work came from the paper by wilczek and rajagopal .
we have benefitted from discussions with many of our colleagues , in particular , v.soni , g.baskaran , deepak kumar , p.k .
mitter j.b .
zuber and e. seiler . 99 for a review see , _ aspects of symmetry _ , s.coleman ( cambridge univ .
press(1985 ) ) . for
a review see , _ current algebra _
s.adler and r.dashen , a. benjamin inc .
( 1968 ) f.wilczek , int.j.mod.phys.*a7*(1992 ) 3911 ; k.rajagopal and f.wilczek , nucl.phys .
* b399 * ( 1993 ) 395 .
s. weinberg , physica * 96a * ( 1979 ) 341 .
for reviews see , h.leutwyler,_recent aspects of quantum fields_,ed .
h.mitter and m.gausterer , lecture notes in physics * 396*(1991)1;_principles of chiral perturbation theory _
, preprint no butp-94/13 , also published in hadron physics 94 , 1 - 46 ; hep - ph/9406283 .
h. lehmann , phys .
* b41 * , ( 1972 ) 529 ; h. lehmann and h. trute , nucl . phys .
* b52 * ( 1973 ) 280 ; l .- f .
li and h. pagels , phys .
* 26 * ( 1971 ) 1204 ; _ ibid . _
* 27 * ( 1971 ) 1089 ; p. langacker and h. pagels phys . rev .
* d8 * ( 1973 ) 4595 . see for instance , j.zinn-justin , _ quantum field theory and critical phenomena _ , clarendon press , oxford , 1989 . t. holstein and h. primakoff , phys .
rev.*58 * , 1098(1940 ) .
wallace and r.k.p .
zia , phys . rev . *
b12*,5340(1975 ) . w.a.bardeen , b.w.lee and r.e.schrock * d14*(1976)985 .
s.weinberg , phys . rev .
lett * 18 * 188(1967 ) t.h berlin and m. kac , phys .
rev * 86 * ( 1952 ) 521 ; h.e .
stanley , phys .
rev * 176*(1968)718 .
a.z . patashinsky , v.l .
pokrovsky , sov phys jetp * 37*,733(1973 ) .
see also , v.g .
vaks , a.i .
larkin and s.i .
pikin , sov .
26*,647(1968 ) ; m.e .
fisher , m.n .
barber and d. jasnow , phys .
rev * a8*,1111(1973 ) .
patashinsky , v.l .
pokrovsky,_fluctuation theory of phase transition _ ,
moscow , nauka 2075:pergamon , london .
e. brezin , d.j .
wallace and k.g .
wilson , phys . rev . *
b7 * , 232(1973 ) .
f.dunlop and c.m.newman , commun .
phys.*44 * ( 1975)223 .
+ j. lebowitz and o. penrose ( as communicated to dunlop and newman ) e. brezin and j.zinn-justin , phys .
rev . lett.*36 * ( 1976 ) 691 .
private communication by j.zinn-justin .
amit , _ field theory , the renormalisation group and critical phenomena _ ,
world scientific , 1980 .
i.ya.arefeva,e.r.nissimov and s.j.pacheva commun .
* 71 * ( 1980 ) 213 ; i.ya.arefeva,s.i.azakov , nucl.phys .
* b162 * ( 1980 ) 298 .
f. karsch , phys .
* d49 * ( 1994 ) 3791 .
+ f. karsch and e. laerman , phys
* d50*(1994 ) 6954 .
+ c. de tar , nucl .
* 42 * ( 1995 ) 457 .
+ koetzler and muschke , phys rev * b34*(1986)3543 ; a.kocic , j.b.kogut and m - p lombardo , nuc phys * b413*(1994 ) 503 . | we develop various complementary concepts and techniques for handling quantum fluctuations of goldstone bosons.we emphasise that one of the consequences of the masslessness of goldstone bosons is that the longitudinal fluctuations also have a diverging susceptibility characterised by an anomalous dimension @xmath0 in space - time dimensions @xmath1.in @xmath2 these fluctuations diverge logarithmically in the infrared region.we show the generality of this phenomenon on the basis of i ) .
renormalization group flows , ii ) .
ward identities , and iii ) .
schwinger - dyson equations.we also obtain an explicit form for the generating functional of one - particle irreducible vertices of the o(n ) ( non)linear @xmath3models in the leading @xmath4 approximation.we show that this incorporates all infrared behaviour correctly both in linear and non - linear @xmath3 models .
these techniques provide an alternative to chiral perturbation theory.some consequences are discussed briefly .
* infrared behaviour of systems with goldstone bosons * + ramesh anishetty , rahul basu , n.d .
hari dass and h.s.sharatchandra + _ the institute of mathematical sciences , madras-600 113 , india _ |
the study of monotonicity and rigidity of solutions to semilinear elliptic equations of fractional order in the whole space @xmath0 or in smooth bounded sets @xmath1 has attracted considerable attention in the last years , see e.g. @xcite . in striking contrast ,
if @xmath1 is unbounded , but different from the whole space , very few results are available , all concerning the particular case of the half - space @xmath2 , see @xcite , or the one of exterior sets , see @xcite .
the main purpose of this paper is the study of the qualitative properties of bounded solutions to @xmath3 where @xmath1 is assumed to be _ the epigraph of a continuous function @xmath4 _ , i.e. we suppose that @xmath5 notice that the half - space @xmath6 falls within this definition with @xmath7 . in , @xmath8 with @xmath9 denotes the fractional laplacian , which can be defined as the operator acting on sufficiently smooth functions as @xmath10 where @xmath11 is a normalizing constant ( which plays no major role in the present paper and which will be often omitted for simplicity ) , and @xmath12 stays for principal value "
. we will consider different assumptions on @xmath13 , obtaining different monotonicity and rigidity properties accordingly .
the nonlinearity @xmath14 in belongs to a reasonably wide class of functions , including for instance those of bistable - type ( a precise definition will follow shortly ) . under these assumptions ,
the main results of this paper are : * boundary regularity , monotonicity and further qualitative properties for solutions to in globally lipschitz epigraphs ; * monotonicity for solutions to in coercive epigraphs ; * @xmath15-dimensional symmetry in the half - space ; * rigidity for overdetermined problems in epigraphs that are sufficiently flat at infinity " .
similar results in the classical case @xmath16 were obtained in the seminal paper @xcite . here
, dealing with a nonlocal framework , a careful analysis is needed to overcome the lack of explicit barriers and several ad - hoc arguments will be exploited to replace the study of the point - wise behavior of the solution with a global study of the geometry of the problem . as additional statements , we also derive a very general maximum principle ( tailor - made for non - decaying solutions in possibly unbounded domains ) , a general version of the sliding method for the fractional laplacian , and a boundary regularity result for solutions of fractional boundary value problems in sets satisfying an exterior cone condition . before proceeding with the statement of our results , we clarify that with the terminology _ solution _ in this paper we always mean _ classical solution_. as a matter of fact , without extra effort , the same results would apply to bounded _ viscosity solutions _ of : indeed , since we will assume that the nonlinearity @xmath14 is locally lipschitz continuous , the regularity theory for viscosity solutions ( developed in @xcite ) implies that viscosity and classical solutions coincide in our setting ( see ( * ? ? ?
* remark 2.3 ) for a detailed explanation ) .
in addition , we mention that _ distributional ( i.e. very weak ) solutions _ or _ weak solutions _ ( as defined e.g. in @xcite ) could be considered as well with minor changes . for the reader s convenience , we will recall the definition of classical and viscosity solution at the end of the introduction . in the next subsections , we describe in details the results obtained . in all the forthcoming statements ,
the fractional parameter @xmath17 will always be a fixed value in the interval @xmath18 .
in this subsection we consider the case in which the domain @xmath1 of is a _ globally lipschitz epigraph_. namely , we suppose that the function @xmath13 in is globally lipschitz continuous , with lipschitz constant @xmath19 . on the nonlinearity @xmath14
, we suppose that : * @xmath14 is locally lipschitz continuous in @xmath20 , and there exists @xmath21 such that @xmath22 for any @xmath23 , and @xmath24 for any @xmath25 ; * there exist @xmath26 , and @xmath27 such that @xmath28 for any @xmath29 $ ] ; * there exists @xmath30 such that @xmath14 is non - increasing in @xmath31 . as prototype example
, we may think at @xmath32 , which yields the fractional allen - cahn equation , that is a widely studied model in phase transitions in media with long - range particle interactions , see e.g. @xcite .
the first of our main results is the natural counterpart of theorems 1.1 and 1.2 in @xcite .
[ thm : main 1 ] let @xmath1 be a globally lipschitz epigraph .
let @xmath14 satisfy assumptions ( @xmath33)-(@xmath34 ) , and let @xmath35 be a bounded solution to . then : * @xmath36 in @xmath1 .
* as @xmath37 , we have that @xmath38 uniformly in @xmath1 .
* there exist @xmath39 such that @xmath40 * @xmath35 is globally @xmath41-hlder continuous in @xmath0 , for some @xmath42 .
* @xmath35 is the unique bounded solution to .
* if @xmath43 is such that @xmath44 then @xmath45 in particular , @xmath35 is monotone increasing in @xmath46 .
we stress that , since @xmath1 is merely a lipschitz set and the exterior sphere condition is not necessarily satisfied along @xmath47 , the hlder continuity of the solution does not follow by previous contributions ( see subsection [ au45678iuhauaj13 ] for more details ) .
both the exponents @xmath48 and @xmath41 appearing in the theorem are determined by the choice of @xmath13 . to be more precise , we note that a globally lipschitz epigraph satisfy both a uniform exterior cone condition with angle @xmath49 and a uniform interior cone condition with angle @xmath50 ( see subsection [ au45678iuhauaj13 ] for a precise definition of the exterior cone condition ) ; then , as it will be clear from the proofs , the exponent @xmath41 depends on @xmath49 , while the index @xmath48 depends on @xmath50 .
regarding point ( @xmath51 ) in the theorem , it establishes that @xmath35 is monotone increasing in any direction @xmath52 such that there exists an orthonormal basis @xmath53 of @xmath0 with @xmath54 , and in the new coordinates @xmath1 is still the epigraph of a globally lipschitz function @xmath55 .
thus , in the particular case @xmath7 , i.e. when @xmath2 is a half - space , we deduce monotonicity and @xmath15-dimensional symmetry of the solutions .
[ cor : half - space ] let @xmath2 , and let @xmath14 satisfy ( @xmath33)-(@xmath34 ) , and let @xmath35 be a bounded solution to . then @xmath35 depends only on @xmath46 , and @xmath56 previous results regarding monotonicity of solutions to nonlocal equations in half - spaces can be found in @xcite ( see also @xcite for results in the whole of @xmath0 ) , where the authors dealt with non - decreasing nonlinearities @xmath14 satisfying @xmath57 .
we also refer to @xcite ( which appeared after the present paper was submitted ) , where the authors proved that any bounded , non - negative and non - trivial solution to with @xmath14 of class @xmath58 is montone increasing . in all the articles @xcite the monotonicity
is used to derive non - existence results for non - decreasing nonlinearities @xmath14 , which is a complementary situation with respect to the one considered here .
the @xmath15-dimensional symmetry in the half - space , as addressed in corollary [ cor : half - space ] , was , up to now , open .
coming back to the monotonicity of the solutions , we emphasize that the main result in @xcite allows to treat also the case @xmath59 .
this marks a relevant difference with the local setting @xmath16 , since in case @xmath59 non - negative solutions of local equations are not necessarily monotone , and only partial results are available ( we refer the interested reader to @xcite ) .
the proof of theorem [ thm : main 1 ] is given in section [ sec : thm 1 ] , and relies on some classical ideas of @xcite nevertheless , all the intermediate steps present several substantial difficulties of purely nonlocal nature . as a matter of fact , in @xcite the authors often construct more or less explicit local barriers , and exploit local properties of functions whose laplacian has a strict sign . on the other hand ,
the construction of a barrier function is much harder when dealing with integro - differential operators , since such barrier has to be defined in the all space @xmath0 , and has to satisfy a boundary condition on the complement of a certain set @xmath60 ( and not only on @xmath61 ) .
moreover , local properties of functions can not be inferred by the only knowledge of the fractional laplacian in some neighbourhood and any modification of the function `` far away '' affects the values of its fractional laplacian at a point .
these are just two sources of new obstructions which we shall overcome ; we refer to the comments and the remarks written throughout the paper for further details . in this subsection
, we deal with the case in which the domain @xmath1 of is a _ coercive epigraph _ , namely , we suppose that the function @xmath62 in is continuous and satisfies @xmath63 in this setting , we have the following result : [ thm : main 2 ] let @xmath1 be a coercive epigraph , and let @xmath35 be a solution ( not necessarily bounded ) to@xmath64 with @xmath65 continuous in @xmath66 , non - decreasing in @xmath46 , and locally lipschitz continuous in @xmath67 , locally uniformly in @xmath68 , in the following sense : for any @xmath69 and any compact set @xmath70 , there exists @xmath71 such that @xmath72.\ ] ] then @xmath35 is monotone increasing in @xmath46 .
this result is the natural counterpart of ( * ? ? ?
* theorem 1.3 ) , which in turn is a refinement of ( * ? ? ?
* proposition ii.1 ) .
its proof rests on the moving planes method for the fractional laplacian . in this subsection
, we consider the overdetermined setting in which both dirichlet and neumann conditions are prescribed in problem . differently from the classical case
, the dirichlet condition needs to be set in the complement of the domain ( and not along its boundary ) and the neumann assumptions takes into account ( in a suitable sense ) normal derivatives of fractional order . for this , given an open set @xmath1 with @xmath73 boundary , we denote by @xmath74 the inner unit normal vector at @xmath75 .
for any @xmath76 and @xmath77 , we consider the _ outer normal @xmath17-derivative _ of @xmath35 in @xmath78 , defined as @xmath79 the boundary regularity theory for fractional laplacian , developed in @xcite , ensures that , for a solution @xmath35 to with @xmath1 of class @xmath73 , the quantity @xmath80 is well defined .
natural hopf s lemmas were then proved in ( * ? ? ?
* proposition 3.3 ) and ( * ? ? ?
* lemma 1.2 ) , and constituted the base point in the study of overdetermined problems for the fractional laplacian , see @xcite .
in this paper we consider overdetermined problems of the type @xmath81 we will suppose that @xmath1 is the epigraph of a @xmath82 and globally lipschitz function @xmath62 , satisfying the following additional assumption : @xmath83 this condition , firstly proposed in @xcite , can be seen as a flatness condition of @xmath47 at infinity
. we can extend ( * ? ? ?
* theorem 7.1 ) in the nonlocal setting .
[ thm : overdet ] let @xmath1 be the epigraph of a @xmath73 and globally lipschitz function @xmath62 , satisfying .
let @xmath14 satisfy ( @xmath33)-(@xmath34 ) , and let us suppose that has a bounded solution @xmath35 .
then @xmath1 is a half - space @xmath84 , and @xmath35 depends only on @xmath46 and is monotone increasing in @xmath46 .
theorem [ thm : overdet ] is proved in section [ sec : overdet ] . in this subsection , we obtain general boundary regularity results for solutions to @xmath85 in this setting , we will not restrict to the case in which @xmath1 is an epigraph , but we will assume instead that @xmath1 satisfies an exterior cone condition with some uniform opening @xmath86 .
more precisely , for a given direction @xmath87 and a given angle @xmath88 , we denote by @xmath89 the open , rotationally symmetric cone of axis @xmath90 and opening @xmath91 ( that is the set of all vectors @xmath92 that form with @xmath90 an angle less than @xmath91 ) .
we suppose that there exists @xmath88 such that : if @xmath93 , then for a direction @xmath94 the cone @xmath95 is exterior to @xmath1 and tangent to @xmath1 in @xmath68 .
we point out that in this definition the opening @xmath91 is the same for all the points of @xmath47 , while the direction @xmath90 can change .
we also observe that globally lipschitz epigraphs @xmath96 enjoy the uniform exterior cone condition , with @xmath91 depending only on the lipschitz constant of @xmath13 .
the boundary regularity of solutions to boundary value problems driven by integro - differential operators has been object of several contributions @xcite .
as far as we know , for boundary value problems of type the minimal assumption on @xmath1 was considered in ( * ? ? ?
* proposition 1.1 ) , where the authors supposed that @xmath1 is a bounded lipschitz domain satisfying a uniform exterior ball condition .
the following theorem weaken this assumption , establishing the global hlder continuity of bounded solutions to when @xmath1 satisfies a uniform exterior cone condition .
[ thm : boundary regularity ] let @xmath1 be a possibly unbounded open set of @xmath0 , satisfying the uniform exterior cone condition with opening @xmath97 .
let @xmath98 be a solution to , with @xmath99 . then
, there exist @xmath42 and @xmath71 , both depending only on @xmath91 , @xmath17 and @xmath100 , such that @xmath101 , and @xmath102.\ ] ] also , for any @xmath9 , the map @xmath103 is monotone non - decreasing .
we stress that it is not even necessary to suppose the lipschitz regularity of @xmath1 .
if @xmath104 , then the uniform exterior cone condition yields a uniform exterior sphere condition , and hence by ( * ? ?
* proposition 1.1 ) solutions to fractional boundary value problems in @xmath1 are already known to be of class @xmath105 .
this is why we only consider @xmath97 in theorem [ thm : boundary regularity ] .
the proof of theorem [ thm : boundary regularity ] is based upon the construction of a wall of barriers , whose definition uses essentially the existence of homogeneous solutions to non - local problems in cones , contained in @xcite .
the homogeneity exponent appearing in this construction coincides precisely with the regularity index @xmath41 , and in particular the monotonicity of @xmath41 in the angle @xmath91 follows from lemma 3.3 in @xcite ( see section [ sec : boundary ] , and in particular remark [ rem : on monotonicity in theta ] , for more details ) . for our purposes ,
the importance of is to provide uniform convergence of sequences of solutions under very reasonable assumptions .
namely , let us consider a sequence @xmath106 of solutions to @xmath107 with @xmath106 and @xmath108 uniformly bounded in @xmath109 and @xmath110 , respectively .
then theorem [ thm : boundary regularity ] implies that @xmath111 locally uniformly in @xmath0 , up to a subsequence .
the proof of theorem [ thm : boundary regularity ] is the object of section [ sec : boundary ] .
we conclude the introduction stating some general results that are auxiliary to the proof of the main theorems and which we think are also of independent interest . in proving theorem [ thm : main 1 ] , a crucial tool will be a maximum principle in unbounded domain for the fractional laplacian .
this is the fractional counterpart of ( * ? ? ?
* theorem 2.1 ) , and we stress that while in the local case the domain @xmath60 is supposed to be connected , this is not necessary in the nonlocal setting .
[ thm : super max ] let @xmath60 be an open set in @xmath0 , possibly unbounded and disconnected .
suppose that @xmath112 is disjoint from the closure of an infinite open connected cone .
let @xmath113 bounded above , and satisfying in viscosity sense @xmath114 for some @xmath115 , @xmath116 a.e . in @xmath60 , with @xmath117 .
then @xmath118 in @xmath60 .
we observe that theorem [ thm : super max ] here improves ( * ? ? ?
* theorem 2.4 ) , where a similar result was proved when @xmath60 was a half - space .
we shall also need the following version of the sliding method for the fractional laplacian .
[ thm : sliding ] let @xmath1 be a bounded open subset of @xmath0 , convex in the direction @xmath119 .
let @xmath120 be a solution of @xmath121 with @xmath122 continuous in @xmath66 , non - decreasing in @xmath46 , and locally lipschitz continuous in @xmath67 , uniformly in @xmath68 , in the following sense : for any @xmath69 , there exists @xmath71 such that @xmath123.\ ] ] on the boundary term @xmath13 , we suppose that for every @xmath124 , @xmath125 , @xmath126 with @xmath127 , it holds @xmath128 then @xmath120 is monotone increasing with respect to @xmath46 . thanks to the maximum principle in sets of small measure (
* proposition 2.2 ) , the proof of theorem [ thm : sliding ] is a straightforward adaptation of the local one , given in @xcite , and hence is omitted . * basic definitions and notations : * we start recalling the definition of classical solution .
a continuous function @xmath129 is a _ classical solution _ to @xmath130 if @xmath131 is well defined and equal to @xmath132 for every @xmath133 , and @xmath134 a.e . in @xmath135 .
notice that , beyond the continuity , no boundary regularity is required on @xmath35 .
it is well known that a sufficient condition to ensure that @xmath136 is well defined ( and actually continuous ) in an open set @xmath137 is that @xmath138 for some @xmath139 ( i.e. @xmath140 if @xmath141 , or @xmath142 if @xmath143 ) , see ( * ? ? ?
* proposition 2.4 ) . for future convenience , we recall here also the definition of viscosity solution ( see ( * ? ? ?
* definition 2.2 ) . a function @xmath144 , upper ( resp .
lower ) semicontinuous is said to be a _ viscosity sub - solution _ ( resp . _ viscosity super - solution _ ) to , if @xmath145 ( resp @xmath146 ) a.e . in @xmath147 , and
if every time all the following happen : * @xmath148 ; * @xmath100 is a neighbourhood of @xmath148 ; * @xmath149 is some @xmath150 function in @xmath100 ; * @xmath151 ; * @xmath152 ( resp .
@xmath153 ) for every @xmath154 ; then if we let @xmath155 we have @xmath156 ( resp .
@xmath157 ) . a function is a viscosity solution if it both a viscosity sub- and super - solution .
we adopt in the rest of the paper a mainly standard notation .
the ball of center @xmath68 and radius @xmath158 is denoted by @xmath159 , and in the frequent case @xmath160 we simply write @xmath161 .
the letter @xmath162 always denotes a positive constant , whose precise value is allowed to change from line to line .
* organization of the paper : * section [ 8iujn9ijn7yhbnikjm ] deals with the maximum principle in unbounded domains , providing the proof of theorem [ thm : super max ] . then , section [ sec : boundary ] is devoted to the boundary regularity in sets satisfying an exterior cone condition , and contains the proof of theorem [ thm : boundary regularity ] .
the monotonicity in globally lipschitz epigraphs and the proof of theorem [ thm : main 1 ] are dealt with in section [ sec : thm 1 ] , while the monotonicity of solutions in coercive epigraphs , with the proof of theorem [ thm : main 2 ] , is the subject of section [ 8uhg6tr4rtgbifdfghjahj ] .
finally , in section [ sec : overdet ] , we consider overdetermined problems and we prove theorem [ thm : overdet ] .
we devote this section to the proof of theorem [ thm : super max ] . to this aim
, we start with some preliminary results .
first of all , we notice that balls centered at boundaries of cones intersect the cones with mass proportional to that of the ball , namely : [ lemma cone 1 ] let @xmath163 $ ] and @xmath164 .
let @xmath165 . then , for any @xmath166 , @xmath167 , depending on @xmath100 and @xmath41 . up to scaling
, we can assume that @xmath168 , so we want to show that , for any @xmath165 , @xmath169 . for a contradiction , we suppose that is false , namely there exists a sequence @xmath170 for which @xmath171 needs to be bounded ( indeed , if @xmath172 and @xmath173 is large enough , then the ball of radius @xmath174 and tangent from the inside to @xmath175 lies in @xmath176 ) .
therefore , up to a subsequence , we have that @xmath177 , for some @xmath178 .
hence , by using the dominated convergence theorem , we can pass to the limit as @xmath179 and obtain that @xmath180 which is a contradiction with the lipschitz regularity of the cone . for a different proof , see @xcite .
here is another auxiliary results concerning the geometry of cones : [ lemma cone 2 ] let @xmath163 $ ] and @xmath181 as in lemma [ lemma cone 1 ] .
let @xmath182 .
let @xmath183 .
then , @xmath184 for some @xmath185 , depending on @xmath100 and @xmath41 .
let @xmath186 be such that @xmath187 we claim that @xmath188 indeed , if @xmath189 then @xmath190 then , and lemma [ lemma cone 1 ] imply the desired result .
we are now in position to complete the proof of theorem [ thm : super max ] .
we consider @xmath191 .
we claim that @xmath192 in the viscosity sense . to prove this , let @xmath149 be a smooth function touching @xmath193 from above at some point @xmath78 .
we have two cases : either @xmath194 or @xmath195 .
suppose first that @xmath194 .
then @xmath196 and @xmath197 .
accordingly , @xmath198 that is , @xmath149 touches @xmath199 from above at @xmath196 .
thus , by , we have that @xmath200 .
now we consider the case in which @xmath195 .
then @xmath201 as a consequence , @xmath149 has a minimum at @xmath78 and therefore @xmath202 .
accordingly , we have that @xmath203 this completes the proof of . from and
the fact that @xmath204 we conclude that @xmath205 in the viscosity sense .
now , in order to complete the proof of theorem [ thm : super max ] , we want to show that @xmath193 vanishes identically .
suppose not : then there exists @xmath206 such that @xmath207 .
then , recalling that @xmath199 is bounded from above , we set @xmath208 and we take a maximizing sequence @xmath209 such that @xmath210 of course , up to neglecting a finite number of indices , we may suppose that @xmath211 .
hence , since , by , we know that @xmath212 outside @xmath60 , we have that @xmath213 .
we denote by @xmath181 a cone that lies outside @xmath60 ( whose existence is warranted by assumption ) .
then we have that @xmath214 , and we set @xmath215 notice also that @xmath216 .
so , by lemma [ lemma cone 2 ] , we know that @xmath217 for some @xmath185 . thus , we are associated to @xmath8 we have to take a positive small number ( see ( * ? ? ?
* section 2 ) ) .
then , for any @xmath218 , it results that @xmath219 i.e. assumption ( 4.9 ) holds . ] in the position of applying ( * ? ? ?
* corollary 4.5 ) for @xmath193 in each ball @xmath220 , and we obtain that @xmath221 in @xmath222 , for some @xmath223 . but this says that @xmath224 and so , taking the limit and recalling , we obtain @xmath225 , which is a contradiction . we conclude this section recalling the strong maximum principle for the fractional laplacian , whose simple proof is omitted for the sake of brevity .
[ prop : strong ] let @xmath226 be an open set , neither necessarily unbounded , nor connected .
let @xmath120 be a classical solution to @xmath227 with @xmath228 and @xmath229 .
then either @xmath230 , or @xmath231 in @xmath0 .
in this section we analyze the global regularity of solutions to the boundary value problem and we prove theorem [ thm : boundary regularity ] . here
we will assume that @xmath232 and @xmath1 is a set satisfying the uniform exterior cone condition with opening @xmath91 , as defined in subsection [ au45678iuhauaj13 ] .
we recall that , for @xmath94 and @xmath233 $ ] , we denote by @xmath89 the open cone of rotation axis @xmath234 and opening @xmath91 . in particular , if @xmath235 , then @xmath236 .
\end{cases}\ ] ] in this framework , theorem [ thm : boundary regularity ] follows as a corollary of the next intermediate statement : [ prop : 4.1 in bcn ] let @xmath237 be a solution of
. then , there exist @xmath42 and @xmath71 depending only on @xmath91 such that @xmath238 for all @xmath239 .
the proof of proposition [ prop : 4.1 in bcn ] is divided into several lemmas .
we produce a wall of upper barriers for @xmath35 , whose construction is inspired by ( * ? ? ?
* lemma 4.1 ) .
major difficulties arise in our setting in order to compute the fractional laplacian of the barriers . in the next lemma
we establish an estimate for @xmath41-homogeneous functions , which can be seen as a one - side counterpart of the classical euler formula .
[ lem : frac euler ] let @xmath240 , @xmath241 , and let @xmath242 be a positive @xmath41-homogeneous function in @xmath243 , non - negative in the whole space @xmath0 .
then , there exists @xmath71 depending on @xmath49 , @xmath50 and @xmath244 such that @xmath245 for any @xmath246 , we define @xmath247 for any @xmath248 , the fact that @xmath244 is @xmath41-homogeneous and the substitution @xmath249 imply that @xmath250 in particular , by taking @xmath251 and @xmath252 , we obtain that , for any @xmath253 , @xmath254 now we set @xmath255 and @xmath256 we have that @xmath257 . we claim that , in fact , @xmath258 indeed , suppose by contradiction that @xmath259 . then , there exists a sequence @xmath260 in @xmath261 such that @xmath262 also , since @xmath263 is pre - compact , up to a subsequence we may assume that @xmath264 as @xmath179 , for some @xmath265 in the closure of @xmath266 .
this , and fatou lemma imply that @xmath267 consequently , @xmath268 for any @xmath269 , that is @xmath244 is constant on @xmath270 and thus @xmath271 now , since @xmath244 is @xmath41-homogeneous , @xmath272 and so , from the positivity assumption on @xmath244 , we deduce that , fixed @xmath273 with @xmath274 , @xmath275 this is in contradiction with and so it proves . hence , using and , for any @xmath276 , @xmath277 as desired .
now , for the sake of simplicity , we suppose that @xmath278 that @xmath279 and that the cone @xmath280 is exterior to @xmath1 . then we take @xmath281 , so that @xmath282 notice that the admissible range of @xmath283 depends only on the opening @xmath91 . setting @xmath284 ,
the complement of @xmath285 is @xmath286 , with @xmath287 .
let us consider the solution to @xmath288 existence and uniqueness of @xmath244 , up to a multiplicative constant , are proved in ( * ? ? ?
* theorem 3.2 ) .
it is also proved that @xmath244 is @xmath41-homogeneous for some @xmath240 depending on @xmath50 and on @xmath17 , and that the exponent @xmath41 is non - increasing in @xmath50 , i.e. non - decreasing in @xmath91 ( see lemma ( * ? ? ? * lemma 3.3 ) ) . since @xmath289 , the cone @xmath290 contains @xmath6 , and by lemma 3.3 and example 3.2 in @xcite we deduce that @xmath291 .
[ rem : on monotonicity in theta ] the homogeneity exponent @xmath41 appearing here is exactly the regularity index @xmath41 in the thesis of theorem [ thm : boundary regularity ] . therefore , as observed above , the monotonicity of @xmath41 with respect to @xmath91 is a direct consequence of lemma 3.3 in @xcite .
interior regularity theory ensures that @xmath292 .
thus , as @xmath293 and @xmath244 is positive , the restriction of @xmath244 on @xmath294 is bounded from below and from above by positive constants . by homogeneity , and recalling that @xmath244 is uniquely determined up to a multiplicative constant , we can suppose that there exists @xmath295 such that @xmath296 notice that in this way the choice of @xmath244 depends only on @xmath91 ( recall that @xmath297 ) . with this notation , we can now prove the following result : [ lem : barrier 1 ] for @xmath298 , let us define @xmath299 then , there exists @xmath71 depending only on @xmath91 such that @xmath300 in light of lemma [ lem : frac euler ] , we can compute ( notice that we omit the constant @xmath301 and the principal value sense to simplify the notation ) @xmath302 for any @xmath303 , with @xmath71 depending only on @xmath91 and @xmath244 .
this gives the desired result .
since @xmath304 , lemma [ lem : barrier 1 ] suggests that @xmath305 can be used to construct a wall of upper barriers for @xmath35 in @xmath306 .
this is indeed exactly what happens in the classical case , but a similar statement does not hold here , due to the nonlocal nature of the problem . indeed , in order to apply the comparison principle , we should control the sign of the function @xmath307 in the whole complement of @xmath306 and not only along its boundary . on the other hand , one can not conclude directly in our case that @xmath308 in the complement of @xmath306 , since by the function @xmath305 is negative if @xmath309 and @xmath310 is very large . to overcome this problem , in what follows
we define a suitable truncation @xmath311 of @xmath305 , being careful enough so that the thesis of lemma [ lem : barrier 1 ] still holds for @xmath311 , and moreover @xmath312 in @xmath313 . to this purpose ,
let us consider the algebraic equation @xmath314 which possesses the two solutions @xmath315 provided that @xmath316 .
we introduce the set @xmath317 notice that , by , @xmath318 . moreover ,
if @xmath298 is sufficiently small , we have that @xmath319 and so , in particular , @xmath320 now we construct the barrier needed for our purposes : [ lem : barrier 2 ] let @xmath305 be the function introduced in lemma [ lem : barrier 1 ] .
let @xmath321 then , there exist @xmath322 small enough and a constant @xmath323 , both depending only on @xmath91 , such that @xmath324 for every @xmath325 $ ] .
first of all , we observe that @xmath326 indeed , if @xmath327 then necessarily @xmath328 the second inequality is satisfied if and only if @xmath329 .
but @xmath330 , so that @xmath329 implies @xmath331 . this shows that the two conditions in can not take place simultaneously , and proves .
now we recall and we write @xmath332 we claim that @xmath333 to this aim , we point out that @xmath334 , thanks to , and so , by , we obtain that @xmath335 in @xmath336 .
so , to complete the proof of , we focus now on the value of @xmath311 on the points of @xmath337 .
for this , we recall , and , and we observe that @xmath338 now we remark that @xmath339 indeed , if @xmath340 , then by we know that @xmath341 , i.e. @xmath342 which proves . from and , we deduce that @xmath343 as a consequence , @xmath344 \cap
\{w_r < 1\ } = \varnothing,\ ] ] and therefore @xmath335 in @xmath337 for every @xmath345 sufficiently small , which completes the proof of . now we focus on the inequality satisfied by the fractional laplacian of @xmath311 in @xmath346 .
we first observe that if @xmath347 , then @xmath348 , thanks to , and so @xmath349 . using this and lemma [ lem : barrier 1 ]
, we obtain that , for any @xmath347 , @xmath350 \\ & { \leqslant}- c |x|^{2\alpha-2s } + \int_{\r^n } \frac{z_r(x)- z_r(y ) -w_r(x ) + w_r(y)}{|x - y|^{n+2s}}\,dy \\ & = - c |x|^{2\alpha-2s } + \int_{d_r } \frac{w_r(y)- z_r(y)}{|x - y|^{n+2s}}\,dy \\ & = - c |x|^{2\alpha-2s } + \int_{d_r } \frac { 1 - z_r(y)}{|x - y|^{n+2s}}\,dy \\ & { \leqslant}- c |x|^{2\alpha-2s } + \int_{\r^n \setminus b_{c_1/r } } \frac{1 + |z_r(y)|}{|x - y|^{n+2s}}\,dy \end{split}\ ] ] where in the last inequality we used with @xmath351 .
notice also that , if @xmath352 , then @xmath353 consequently , by , latexmath:[\[\label{332 } furthermore , since @xmath298 is appropriately small , we have @xmath355 for any @xmath356 .
hence , recalling that @xmath298 is sufficiently small , that @xmath357 , and that @xmath358 , we obtain that @xmath359 , for any @xmath356 , for an appropriate @xmath323 , as desired .
now we are in the position of completing the proof of proposition [ prop : 4.1 in bcn ] .
let @xmath360 notice that @xmath361 let also @xmath311 be the function introduced in lemma [ lem : barrier 2 ] .
we claim that @xmath362 for @xmath363 $ ] ( where @xmath364 is small enough , as given by lemma [ lem : barrier 2 ] ) . to prove , we observe that , by lemma [ lem : barrier 2 ] and , we know that @xmath365 also , we set @xmath366 notice in particular that @xmath367 . hence , using again lemma [ lem : barrier 2 ] and , we have that @xmath368 for any @xmath369 . this , and the maximum principle ,
imply that @xmath370 in @xmath306 , i.e. @xmath371 for any @xmath369 . on the other hand ,
if @xmath369 , we deduce from that @xmath372 , and therefore , by and , @xmath373 by inserting this into , we conclude that @xmath374 for any @xmath369 .
now we observe that , by , @xmath375 with @xmath162 depending only on @xmath91 ( recall that @xmath376 depends only on @xmath91 ) .
therefore , recalling and , we conclude that @xmath377 for any @xmath369 , for some positive constant @xmath162 depending only on @xmath91 .
the same argument can be repeated replacing @xmath35 with @xmath378 , and so we conclude that , for any @xmath369 , @xmath379 now , take @xmath380 .
we distinguish two cases , either @xmath381 or @xmath382 .
if @xmath381 , we consider @xmath383 to be a projection of @xmath175 along @xmath47 . up to a rigid motion , we may also suppose that @xmath384 ( hence we are in the normalized setting of ) . in this way
, we have that @xmath385 hence , from , we have that @xmath386 which establishes in this case . if instead @xmath382 , we have that @xmath387 then , the estimate in follows in this case from .
this concludes the proof of proposition [ prop : 4.1 in bcn ] .
thanks to the above results , we can now complete the proof of theorem [ thm : boundary regularity ] : since @xmath98 , we can restrict to the case when @xmath388 is small , say @xmath389 .
we prove the thesis of the theorem according to four different cases .
_ case 1 ) _ assume that @xmath390 in this case , we have that @xmath391 .
in particular , by , @xmath392 therefore , by scaling the interior hlder estimate in ( * ? ? ? * proposition 2.9 ) , we deduce that for a constant @xmath71 depending only on @xmath91 @xmath393 |x - y|^\alpha .\ ] ] now , using , we obtain @xmath394 |x - y|^\alpha,\ ] ] which is the desired result in this case .
_ case 2 ) _ now assume that @xmath395 then , by proposition [ prop : 4.1 in bcn ] , @xmath396 as desired .
_ case 3 ) _ now we suppose that @xmath397 and @xmath398 .
we remark that if @xmath399 then @xmath400 hence , for any @xmath399 , we obtain from that @xmath401 accordingly , using the fact that @xmath304 , we obtain that , for any @xmath399 , @xmath402 , we have that @xmath403 for any @xmath404 , up to renaming @xmath71 . as a consequence , @xmath405 estimates and allow us to apply corollary 2.5 in @xcite , which implies that @xmath406 \ , \rho^\alpha.\ ] ] now we observe that @xmath407 and so @xmath408 as a consequence , we deduce from that @xmath409\,\rho^\alpha \ , \left ( \frac{|x - y|}{\rho}\right)^\alpha,\end{aligned}\ ] ] which gives the desired result in this case .
_ case 4 ) _ finally , we consider the case in which @xmath410 by proposition [ prop : 4.1 in bcn ] @xmath411 since in this case @xmath412 , this completes the proof .
in this section , we give the proof of theorem [ thm : main 1 ] . differently from the strategy adopted in @xcite , in our case property ( @xmath413 )
is a direct consequence of the general result in theorem [ thm : boundary regularity ] .
the organization of this section is the following . in subsection
[ sec : convergence ] we prove properties ( @xmath414 ) and ( @xmath415 ) in theorem [ thm : main 1 ] . property ( @xmath416 ) is the object of subsection [ sec : growth ] .
the uniqueness and the monotonicity in @xmath46 are proved in subsection [ sec : uniqueness ] with a unified approach , simplifying the proof in @xcite .
finally , the general monotonicity property in point ( @xmath51 ) of theorem [ thm : main 1 ] follows simply by a suitable rotation of coordinates .
we point out that , for such argument , the global lipschitz continuity of the epigraph @xmath1 is needed . before proceeding , we observe that it is not restrictive to suppose from now on that @xmath417 for the sake of simplicity .
moreover , we define @xmath418 in this way , we have that @xmath419 in @xmath1 . accordingly , in only the restriction of @xmath14 on the interval @xmath420 $ ] plays a role .
therefore , we can modify the definition of @xmath14 outside @xmath420 $ ] without changing the equation , and as a consequence from now on we can suppose that @xmath421 and we denote by @xmath422 its lipschitz constant .
our goal now is to show that @xmath423 in @xmath1 , as claimed in theorem [ thm : main 1]-(@xmath414 ) ( and hence , comparing with , it follows that @xmath424 ) . in order to show that @xmath425 we can proceed as in @xcite , using theorem [ thm : super max ] . for the strict inequality it is sufficient to observe that , being @xmath426 , we are in position to apply the strong maximum principle to the function @xmath427
. for further details , compare with @xcite .
in the rest of the subsection we prove property ( @xmath415 ) in theorem [ thm : main 1 ] ( the local counterpart of this strategy is contained in section 3 in @xcite ) .
[ lem : 3.1 bcn ] let @xmath60 be an open subset of @xmath0 , and let @xmath428 be a locally lipschitz continuous function .
let @xmath244 be a classical supersolution to @xmath429 let also @xmath430 be a ball with closure @xmath431 contained in @xmath60 , and let @xmath199 be a classical subsolution to @xmath432 then for any one - parameter continuous family of euclidean motions @xmath433 with @xmath434 and @xmath435 for every @xmath67 , it results that @xmath436 for every @xmath437 $ ] .
the proof of lemma [ lem : 3.1 bcn ] is very similar to the one of ( * ? ? ?
* lemma 3.1 ) and thus is omitted .
it uses the fact that @xmath8 is invariant under rigid motions and the strong maximum principle . exploiting lemma [ lem : 3.1 bcn ] we can deduce a lower estimate for @xmath35 far away from the boundary @xmath438 .
[ lem : 3.2 bcn ] there exist
@xmath439 , @xmath440 with @xmath441 depending only on @xmath100 and @xmath442 ( recall assumption ( @xmath443 ) ) such that @xmath444 the proof of lemma [ lem : 3.2 bcn ] here is a simple extension to that in ( * ? ? ?
* lemma 3.2 ) and therefore is omitted .
we stress that if @xmath445 denotes the first eigenvalue of @xmath8 in @xmath446 with homogeneous dirichlet boundary condition , then @xmath447 by scaling .
we are now ready to prove the counterpart of lemma 3.3 in @xcite .
we observe that in this framework the local " argument can not be directly extended , since the proof of lemma 3.3 in @xcite heavily relies on local properties of functions whose laplacian has a sign , and this technique does not work in a nonlocal setting .
therefore , to overcome such difficulty , we have to modify the approach in the following way .
[ lem : 3.3 bcn ] let @xmath439 , @xmath440 be given by lemma [ lem : 3.2 bcn ] .
let @xmath448 with @xmath449 , so that @xmath450 .
let @xmath139 so small that @xmath451 .
let @xmath452 \right\}>0.\ ] ] then , there exists @xmath453 depending only on @xmath17 and @xmath100 such that @xmath454^{-2s}.\ ] ] notice that the existence of @xmath455 is guaranteed by the strict upper estimate @xmath456 in @xmath1 , thanks to theorem [ thm : main 1]-(@xmath414 ) .
the crucial fact in the lemma is that @xmath457 does not depend on @xmath458 and @xmath455 .
let @xmath244 be the solution to @xmath459 we claim that the thesis is true with @xmath460 .
if not , we have @xmath461^{-2s},\ ] ] and hence there exists @xmath298 such that @xmath462^{-2s}.\ ] ] notice in particular that @xmath463 , and so @xmath464 let @xmath465 then , by scaling , we have that @xmath466 and the maximum of @xmath199 is @xmath467 .
by , we are now in the position of exploiting lemma [ lem : 3.2 bcn ] : in this way , we have @xmath468 in @xmath469 provided @xmath470 is sufficiently small
. now we increase @xmath52 till we obtain a touching point , namely we let @xmath471 as a matter of fact , since @xmath472 on @xmath473 , we have that @xmath78 lies in the interior of the ball @xmath474 . by definition , and since @xmath199 is radial and radially decreasing with respect to @xmath458 , @xmath475 using this we infer that @xmath476 , and moreover @xmath477 , which together with implies that @xmath478 .
we are ready to complete the contradiction argument : since @xmath476 , by continuity there exists a neighbourhood @xmath479 of @xmath78 such that @xmath480 also , by lemma [ lem : 3.2 bcn ] and , we have that @xmath481 and so , by continuity , we have that @xmath482 thus , in @xmath483 we have that both and are satisfied .
therefore , by the definition of @xmath484 , @xmath485 as a consequence , since @xmath478 , @xmath486 which by the strong maximum principle implies that @xmath487 .
this is a contradiction with the fact that @xmath488 on @xmath489 .
let us assume by contradiction that there exist a sequence of points @xmath490 and some @xmath491 such that @xmath492.\ ] ] then we can choose @xmath139 independent of @xmath493 such that @xmath494 in lemma [ lem : 3.3 bcn ] , deducing that @xmath495 } f { \leqslant}c_1^{-1 } \left [ \dist(y_n,\partial \omega)-r_0 \right]^{-2s } \to 0\ ] ] as @xmath496 .
on the other hand , by assumption ( @xmath33 ) , @xmath497 } f { \geqslant}\min_{[\eps_1,(1+\eps ) ( 1-\rho ) ] } f > 0\ ] ] and this is in contradiction with . in this subsection
we describe the growth of the solution near the boundary of @xmath1 , proving point ( @xmath416 ) of theorem [ thm : main 1 ] .
we point out that the proof is completely different with respect to the one in @xcite ( proof of assertion ( c ) in theorem 1.2 ) , where an argument based on the construction of an explicit local barrier is used .
dealing with a nonlocal operator , such an approach fails , and we should produce a global barrier on which explicit computations are much more involved . to overcome the problem , we replace the barrier - argument with a convenient geometric constructions , which permits to apply iteratively the harnack inequality .
this approach seems to be applicable to a wide class of operators . up to a translation , it is not restrictive to suppose that @xmath498 .
we start with three simple consequences of the fact that @xmath499 is a globally lipschitz epigraph ( recall that @xmath19 denotes the lipschitz constant of @xmath13 ) .
\1 ) let @xmath500 be an infinite open cone , with @xmath501 .
it is possible to choose @xmath502 , depending only on the lipschitz constant @xmath19 , so that if the vertex of @xmath503 is translated to any point of @xmath47 , then the cone is included in @xmath1 : that is there exists @xmath504 such that @xmath505 in what follows we fix @xmath506 .
\2 ) by the lipschitz continuity of @xmath13 , and recalling that @xmath498 ( so that @xmath507 ) , from lemma [ lem : 3.2 bcn ] and theorem [ thm : main 1]-(@xmath415 ) it follows the existence of @xmath508 such that @xmath509 here we implicitly suppose that @xmath510 ; if this were not the case , we can simply replace @xmath439 with a smaller quantity . if necessary
replacing @xmath511 with a larger quantity , we can also suppose that @xmath512 \3 ) finally we observe that @xmath513 and @xmath514 is bounded by construction .
we consider a point @xmath515 let @xmath60 be a bounded domain of @xmath0 such that @xmath516 with @xmath517 sufficiently smooth , see figure [ 12345678 ] . in the proof of theorem [ thm : main 1]-(@xmath416).,width=445 ]
we introduce a continuous function @xmath244 satisfying @xmath518 by the maximum principle , @xmath519 in @xmath60 , and hence by compactness there exists @xmath520 such that @xmath521 also , by , we have that @xmath522 in @xmath523 .
furthermore , @xmath524 in @xmath525 , thanks to assumption ( @xmath33 ) and theorem [ thm : main 1]-(@xmath414 ) . therefore , by the comparison principle in theorem [ thm : super max ] , we find that @xmath526 recalling , let now @xmath527 , and let us define @xmath528 see figure [ figi2 ] .
-(@xmath416).,width=445 ] in this way @xmath529 for every @xmath530 , and @xmath531 as a consequence @xmath532 for every @xmath533 .
we claim that @xmath534 thanks to and , and observing that by construction @xmath535 for every @xmath530 , we see that @xmath536 has to be the minimum natural number satisfying @xmath537 that is @xmath538 notice that such a @xmath536 does exist , thanks to .
this proves .
furthermore , we remark that @xmath539 for suitable @xmath540 , and @xmath541 , all depending only on @xmath542 , @xmath511 and @xmath502 .
this and say that @xmath543 the previous construction implies that we have finite sequences of points @xmath544 and radii @xmath545 , with @xmath546 , such that @xmath547 for every @xmath546 , and @xmath548 with @xmath549 independent of @xmath530 . from and , we deduce that @xmath244 is non - negative in @xmath0 and @xmath17-harmonic in @xmath550 , for any @xmath551 .
hence , we are in position to apply the harnack inequality in ( * ? ? ?
* theorem 5.1 ) , deducing that @xmath552 for every @xmath553 , where @xmath554 is a positive constant depending only on @xmath502 .
now we observe that @xmath555 thanks to and .
similarly , @xmath556 thanks to and . as a consequence of this and of
, we obtain that @xmath557 .
therefore , using and , we obtain @xmath558 for some positive constants @xmath559 depending only on @xmath502 , @xmath19 , @xmath542 and @xmath511 .
this gives the desired point - wise estimate at the point @xmath78 , with @xmath560 and @xmath561 ( recall ) , and with constants @xmath162 and @xmath48 depending only on @xmath502 , @xmath19 , @xmath542 and @xmath511 .
up to a translation , the same estimate holds at any point of @xmath1 , with vertical coordinate @xmath562 .
that is , translating @xmath503 ( and hence @xmath60 ) along the boundary @xmath47 , the family @xmath563 gives a wall of lower barriers for @xmath35 , providing the desired lower estimate at any point of @xmath564 , which completes the proof of theorem [ thm : main 1]-(@xmath416 ) .
this section is devoted to the proof of point ( @xmath244 ) in theorem [ thm : main 1 ] .
our approach is inspired by ( * ? ? ?
* section 5 ) , but we modify it in such a way that the proof gives essentially in one shot both uniqueness and monotonicity of the solution .
this simplifies the argument in @xcite , and completes the proof of theorem [ thm : main 1 ] .
we start with the preliminary observation that , under our assumptions , solutions are bounded away from @xmath15 if the distance from the boundary is finite ( recall ) .
[ lem : 3.4 bcn ] for @xmath565 , let @xmath566 then , any bounded solution to is bounded away from @xmath15 in @xmath567 , namely @xmath568 let @xmath35 be a solution of , and let us suppose by contradiction that @xmath569 as @xmath570 along a sequence @xmath571 .
let us consider then the sequence of translated functions @xmath572 .
we observe that @xmath573 and @xmath574 where @xmath575 since @xmath13 is globally lipschitz continuous , the translated epigraphs @xmath576 converge , up to subsequence , to a limit epigraph @xmath577 , which is still globally lipschitz continuous .
notice also that @xmath576 satisfies a uniform cone condition , with opening of the cone independent of @xmath493 .
thus , by theorem [ thm : boundary regularity ] , the sequence @xmath106 is uniformly bounded in @xmath578 , and hence converges locally uniformly in @xmath0 to a limit function @xmath579 .
thus , by and the stability property of viscosity solutions ( * ? ? ?
* lemma 4.5 ) , we infer that @xmath580 is a viscosity solution to @xmath581 actually , arguing as in ( * ? ? ?
* remark 2.3 ) , we see that @xmath580 is a classical solution .
since also the function constantly equal to @xmath15 is a solution of the equation in ( recall that @xmath426 ) , the strong maximum principle implies that @xmath582 in @xmath0 .
this is in contradiction with the fact that @xmath583 , which follows from . in order to prove theorem [ thm : main 1]-(@xmath244 ) ,
let us consider two bounded solutions @xmath35 and @xmath244 of .
we show that necessarily @xmath522 . exchanging the role of @xmath35 and @xmath244
we deduce also that @xmath584 , whence uniqueness follows . for this , first of all we observe that , by theorem [ thm : main 1]-(@xmath415 ) , there exists @xmath585 such that @xmath586 where @xmath587 was introduced in assumption ( @xmath34 ) .
let @xmath588 also , for @xmath589 let us consider @xmath590 as in ( * ? ? ?
* lemma 5.1 ) , we show that : [ lem : 5.1 bcn ] if @xmath591 in @xmath592 , then @xmath591 in @xmath0 . first of all , since @xmath1 is an epigraph we have @xmath593 now we notice that @xmath594 in @xmath595 .
thus , to establish the desired result , we have only to prove that @xmath596 to this aim , we use and we observe that @xmath597 where @xmath598 now we claim that @xmath599 indeed , let @xmath600
. then @xmath601 , thanks to . hence follows from .
now , since @xmath14 is non - increasing in @xmath602 , we deduce from that @xmath603 in @xmath604 , and by lipschitz continuity we have also @xmath605 .
moreover , it is clear that @xmath604 satisfies an exterior cone condition ( since @xmath1 does ) , and hence theorem [ thm : super max ] and the strong maximum principle imply , as desired .
now , we aim at showing that @xmath606 in @xmath0 for @xmath607 .
thanks to lemma [ lem : 5.1 bcn ] , this statement is equivalent to showing that @xmath606 in @xmath592 for @xmath608 . by lemma [ lem : 3.4 bcn ] , we know that @xmath609 with @xmath610 in @xmath592
. moreover , since @xmath611 uniformly as @xmath37 ( recall theorem [ thm : main 1]-(@xmath415 ) ) , we have that @xmath606 in @xmath592 for @xmath52 sufficiently large .
therefore , we can define @xmath612 [ rem : uniq and monot ] for the uniqueness , one could replace the previous definition of @xmath613 with @xmath614 as done in @xcite . nevertheless , as we will show later , definition permits to perform the same argument used for the uniqueness also for the monotonicity of @xmath35 .
we are now in the position of completing the proof of theorem [ thm : main 1]-(@xmath244 ) . by continuity @xmath615 in @xmath592 .
hence , by lemma [ lem : 5.1 bcn ] , @xmath616 thus , if @xmath617 the proof is complete . to rule out the possibility that @xmath618 , we argue by contradiction . if @xmath618 , then there exist sequences @xmath619 and @xmath620 such that @xmath621 and @xmath622 let us consider @xmath623 as in lemma [ lem : 3.4 bcn ] , we have that , up to subsequences , @xmath624 and @xmath625 locally uniformly , and @xmath580 and @xmath626 are solutions to in a limit epigraph @xmath577 .
we remark that , since @xmath627 , the point @xmath628 belongs to the approximating domains and therefore @xmath629 we also notice that @xmath630 thanks to .
furthermore , in light of , and using the uniform convergence , @xmath631 by and , we conclude that @xmath632 now we claim that @xmath633 to check this , we take @xmath634
. then there exists @xmath635 such that @xmath636 as @xmath637 . as a consequence , using the uniform convergence , we see that @xmath638 and @xmath639 since @xmath640 ( here we are using that @xmath618 ) . combining and , we obtain . also , @xmath641 where @xmath642 and @xmath643 .
thus , the strong maximum principle and imply that either @xmath644 in @xmath577 , or @xmath645 in @xmath0 .
but the latter alternative is not admissible , due to .
therefore , we conclude that @xmath644 in @xmath577 and so , again by , in the closure of @xmath577 .
this is in contradiction with and , hence the proof of theorem [ thm : main 1]-(@xmath244 ) is complete .
we proceed exactly as for the uniqueness , but instead of comparing @xmath646 with @xmath244 we compare @xmath646 with @xmath35 ( indeed , @xmath244 was just the generic solution , so the case @xmath647 is admissible ) . in the end , we obtain that @xmath648 in @xmath592 for every @xmath649 , which by lemma [ lem : 5.1 bcn ] yields the desired monotonicity .
this section is devoted to the proof of theorem [ thm : main 2 ] , which rests upon the moving planes method .
we introduce some notation : for @xmath650 , we set @xmath651 the crucial remark is that , since we deal with a coercive epigraph , the set @xmath652 is bounded for every @xmath650 , even if @xmath1 is unbounded . therefore , one can adapt the proof of ( * ? ? ?
* theorem 1.1 ) , which uses the moving planes method for fractional elliptic equations in bounded domains . for the reader
s convenience , we recall the following weak maximum principle in sets of small measure , which we conveniently re - phrase for our purpose . [
lem : small ] let @xmath60 be an open and bounded subset of @xmath0 .
let @xmath115 with @xmath653 , and let @xmath199 be a solution to @xmath654 then , there exists @xmath185 depending only on @xmath100 , @xmath17 and @xmath655 such that if @xmath656 then @xmath657 in @xmath60 .
[ rem : uniform delta ] consider a sequence of boundary value problems of type , with @xmath658 and @xmath659 , @xmath660 .
if we have a uniform bound @xmath661 , then proposition [ lem : small ] gives a threshold @xmath662 independent of @xmath493 .
we set @xmath663 and @xmath664 .
we aim at proving that @xmath665 in @xmath666 for every @xmath667 , which gives the desired monotonicity . for any @xmath667
, we have that @xmath668 in @xmath652 .
accordingly , the monotonicity of @xmath14 in @xmath46 gives @xmath669 in @xmath652 , with @xmath670 notice that , thanks to the lipschitz continuity of @xmath14 and the fact that @xmath671 , for any @xmath672 , there exists @xmath71 such that @xmath673 for any @xmath674 $ ] . for convenience ,
we now divide the proof into separate steps : _ step 1 ) we show that @xmath665 in @xmath652 for any @xmath667 , with @xmath675 small enough . _ for this
, let @xmath676 .
we first show that @xmath677 i.e. , that @xmath678 . to this aim , we argue by contradiction .
if @xmath679 , we can define @xmath680 we observe that @xmath681 , and that @xmath682 while @xmath683 .
exactly as in ( * ? ? ? * theorem 1.1 , step 1 ) , it is possible to show that @xmath684 in @xmath685 .
hence , by , @xmath686 thus , by the maximum principle in sets of small measure ( see proposition [ lem : small ] ) , we infer that , for any @xmath687 close to @xmath688 , it results that @xmath689 in @xmath0 . as a consequence , @xmath690 in @xmath691 , which proves . as a side remark
, we notice that here we do not need @xmath98 , but only @xmath692 ( which follows automatically by the definition of classical or even viscosity solution ) , since @xmath652 is bounded .
now we claim that @xmath693 this is not a consequence of the strong maximum principle since the function @xmath694 changes sign , by definition . by contradiction ,
let us suppose that there exists @xmath695 such that @xmath696 . since @xmath697 for every @xmath695 and @xmath698 , and @xmath694 is positive in a subset of @xmath666 having positive measure , we deduce that @xmath699 where was used , and so we obtain a contradiction .
this proves .
then , the desired result in step 1 follows by combining and .
_ step 2 ) we show that @xmath700 in @xmath652 for every @xmath687 .
_ let @xmath701 by the previous step @xmath702 . if @xmath703 the proof of step 2 is complete , and hence we argue by contradiction supposing that @xmath704
. by continuity and by we have @xmath705 let us consider now @xmath706 } \|c_\lambda\|_{l^\infty(\sigma_\lambda)}.\ ] ] this value is finite since @xmath671 , @xmath14 is locally lipschitz , and @xmath707 is bounded .
therefore , the threshold @xmath708 for the maximum principle in domains of small measure in proposition [ lem : small ] is well defined ( recall remark [ rem : uniform delta ] ) .
let us fix a compact set @xmath709 such that @xmath710 using this and , we have that , by continuity , there exists @xmath711 small enough such that @xmath712 for every @xmath713 .
let now @xmath714 .
we observe that @xmath715 thanks to . now
, we consider the functions @xmath716 and @xmath717 defined as in with @xmath718 .
proceeding as in step 1 , we can check that @xmath719 we use this , and the maximum principle in sets of small measure ( see proposition [ lem : small ] ) to conclude that @xmath720 in @xmath0 for every @xmath455 sufficiently small . as a consequence , recalling , we have that @xmath721 . therefore , from
we conclude that @xmath722 in @xmath723 for any @xmath139 small enough , in contradiction with the definition of @xmath724 .
this section concerns the study of the overdetermined problem , where @xmath1 is a globally lipschitz epigraph , satisfying the additional flatness assumption .
regarding @xmath14 , it satisfies ( @xmath33)-(@xmath34 ) in the introduction . as in the proof of theorem [ thm : main 1 ]
, it is not not restrictive to suppose that @xmath725 .
in particular , we now proceed with the proof of theorem [ thm : overdet ] , which is the fractional counterpart of the proof of theorem 7.1 in @xcite , where the local case was considered .
we claim that for any @xmath726 @xmath727 to this extent , for @xmath728 fixed and @xmath729 , let us consider @xmath730 since @xmath13 is lipschitz continuous , for @xmath731 sufficiently large we have that @xmath732 contains strictly @xmath1 .
in other words , we can define the real number @xmath733 we claim that @xmath734 to prove this , let us suppose by contradiction that @xmath735 . then there exist sequences @xmath736 and @xmath737 with @xmath738 by assumption
, we infer that @xmath739 is bounded , and hence up to a subsequence @xmath740 , for some @xmath741 . on the other hand , by , we know that @xmath742 and therefore @xmath743 . in other words ,
the set @xmath1 is internally tangent to @xmath744 in @xmath745 .
now , let us consider , for @xmath746 , @xmath747 we claim that @xmath748 to this aim , we argue as in subsection [ sec : uniqueness ] .
first , we introduce @xmath585 large enough , so that both @xmath35 and @xmath749 are larger than @xmath750 in @xmath751 ( @xmath750 is defined in assumption ( @xmath34 ) ) .
then , for any @xmath746 , we have that @xmath752 in @xmath604 , and @xmath753 in @xmath595 . by lemma [ lem : 3.4 bcn ] and theorem [ thm : main 1 ] applied to @xmath35 , we know that @xmath754 in @xmath755 if @xmath756 is sufficiently large
. therefore we can define @xmath757 if @xmath758 , then claim follows from lemma [ lem : 5.1 bcn ] . on the other hand , if @xmath759 it is not difficult to obtain a contradiction as in subsection [ sec : uniqueness ] , thus completing the proof of .
moreover , by internal tangency , the outer normal to @xmath47 and to @xmath760 coincide at the point @xmath745 .
accordingly , by the @xmath17-neumann condition in reads @xmath761 on the other hand , the function @xmath762 satisfies @xmath763 where @xmath764 therefore , the hopf lemma for the fractional laplacian ( see ( * ? ? ?
* lemma 1.2 ) ) gives @xmath765 in contradiction with .
this proves . using
, we deduce that @xmath766 which in turn implies .
now , we deduce from that the function @xmath13 is necessarily a constant , i.e. @xmath1 is a half - space ( and this concludes the proof of theorem [ thm : overdet ] ) . indeed , if by contradiction @xmath13 is not constant there exist @xmath767 such that @xmath768 .
let @xmath769 . notice that @xmath770 , and therefore @xmath771 thus , using with @xmath772
, we obtain that @xmath773 by adding @xmath774 to this inclusion , we find that @xmath775 , which is in contradiction with .
in a preliminary version of this paper ( see @xcite ) , the proof of lemma [ lem : frac euler ] was unnecessarily complicated : we are indebted to mouhamed moustapha fall for the simpler argument that we incorporated in the present version of this paper . this work has been supported by the alexander von humboldt foundation , the erc grant 277749 _ e.p.s.i.l.o.n . _ `` elliptic pde s and symmetry of interfaces and layers for odd nonlinearities '' , the prin grant 201274fyk7 `` aspetti variazionali e perturbativi nei problemi differenziali nonlineari '' and the erc grant 339958 _ com.pat .
_ `` complex patterns for strongly interacting dynamical systems '' .
s. dipierro , n. soave , and e. valdinoci . on fractional elliptic equations in lipschitz sets and epigraphs : regularity ,
monotonicity and rigidity results .
preprint http://www.wias-berlin.de/preprint/2256/wias_preprints_2256.pdf , 2016 . | we consider a nonlocal equation set in an unbounded domain with the epigraph property .
we prove symmetry , monotonicity and rigidity results .
in particular , we deal with halfspaces , coercive epigraphs and epigraphs that are flat at infinity .
these results can be seen as the nonlocal counterpart of the celebrated article @xcite . |
we start by showing that the sequence @xmath43 is sub additive ; the methodology is similar to gallager s proof in ( * ? ? ?
9.8.1 ) for the case of no feed - forward . then , by showing that the sequence @xmath4 is sub - additive , following ( * ? ? ?
* lemma 4a.2 ) we obtain our main objective , i.e. , @xmath295 to commence , we recall that a sequence @xmath213 is called sub - additive if for all @xmath214 , @xmath215 let @xmath296 be arbitrary positive integers and , for a given @xmath23 , let @xmath297 and @xmath298 be the conditional pmfs that achieve the minimum of the directed information with block length of @xmath1 and @xmath299 , i.e. , that achieve @xmath4 and @xmath300 , respectively .
suppose we transmit @xmath301 samples as follows ; the first @xmath1 samples are transmitted using @xmath302 , and the sequential @xmath299 samples are transmitted using @xmath303 .
hence , the overall conditional pmf is @xmath304 we can see in section [ secalgs ] that the directed information can be written as @xmath305 from the construction of the conditional overall pmf @xmath306 , its clear that @xmath307 furthermore , @xmath308 thus , it follows that @xmath309 since the source is stationary , we can start the input block at any given time index ; thus the pmfs @xmath302 and @xmath303 achieve @xmath310 on the right - hand side of equation ( [ upb ] ) , while the left - hand side is greater than @xmath311 since we attempt to minimize the expression to achieve the rate distortion function .
hence , we obtain @xmath312 using ( * ? ? ?
* lemma 4a.2 ) for sub - additive sequences , we obtain @xmath313
in this appendix we prove lemma [ lemconverse ] , which provides for us that the mathematical expression for the rate distortion feed - forward @xmath211}\leq d}i(\hat{x}^n\to x^n),\label{rdi2}\end{aligned}\ ] ] is a lower bound to the operational definition @xmath35 .
consider any @xmath61 rate distortion with feed - forward code defined by the mappings @xmath314 as given in section [ secprobres ] , equation ( [ mapenddec ] ) , and distortion constraint @xmath3}\leq d+\epsilon_n$ ] , where @xmath315 as @xmath1 goes to infinity .
let the message sent be a random variable @xmath316 , and assume that the distortion constraint is satisfied .
then we have the following chain of inequalities : @xmath317 where ( a ) follows from the fact that the alphabet of @xmath26 is @xmath318 , ( b ) follows from the chain rule for mutual information , ( c ) is due to the fact that given @xmath319 , we know @xmath320 , and ( d ) is since conditioning reduces the entropy .
step ( e ) follows the chain rule for directed information .
taking @xmath1 to infinity , we obtain @xmath321 , and the distortion constraint satisfies @xmath71}\leq d.\nonumber\end{aligned}\ ] ]
in this appendix we provide a proof for theorem [ thrdmax ] .
we recall that theorem [ thrdmax ] states that the rate distortion function can be written as the following optimization problem : @xmath322 where , for some causal conditioned probability @xmath217 , @xmath218 satisfies the inequality constraint @xmath323 we prove this theorem in two ways .
one is similar to berger s proof in @xcite , based on the inequality @xmath220 , for the regular rate distortion function .
the other is using the lagrange duality between the minimization problem we are familiar with and a maximization problem as presented in @xcite and @xcite .
we also provide the connection between the curve of @xmath4 and the parameter @xmath221 ; this is embodied in lemma [ rtag ] .
before we begin , we recall that a step in alg .
[ algs ] is defined by the following equality @xmath324 this equality is the outcome of differentiating the lagrangian when @xmath325 is fixed , as given in section [ secder ] .
we shall use this equality throughout the proof .
as mentioned , the first proof follows the one in @xcite .
first , we show that for every @xmath326 for which the distortion constraint is satisfied , the following chain of inequalities holds @xmath327}-\sum_{x^n}p(x^n)\log\gamma(x^n)\nonumber\\ & = \sum_{x^n,\hat{x}^n}p(x^n)r(\hat{x}^n|x^n)\log\frac{r(\hat{x}^n|x^n)2^{\lambda d(x^n,\hat{x}^n)}}{q(\hat{x}^n||x^{n-1})\gamma(x^n)}\nonumber\\ & \stackrel{(b)}{\geq}\sum_{x^n,\hat{x}^n}p(x^n)r(\hat{x}^n|x^n ) \left(1-\frac{q(\hat{x}^n||x^{n-1})\gamma(x^n)}{r(\hat{x}^n|x^n)2^{\lambda d(x^n,\hat{x}^n)}}\right)\nonumber\\ & = 1-\sum_{x^n,\hat{x}^n}q(\hat{x}^n||x^{n-1})p(x^n)\gamma(x^n)2^{-\lambda d(x^n,\hat{x}^n)}\nonumber\\ & \stackrel{(c)}{\geq}1-\sum_{x^n,\hat{x}^n}q(\hat{x}^n||x^{n-1})p'(x^n||\hat{x}^n)\nonumber\\ & \stackrel{(d)}{=}0,\nonumber\end{aligned}\ ] ] where ( a ) follows from the fact that the distortion @xmath23 exceeds @xmath328}$ ] for every @xmath326 as has been assumed , ( b ) follows from the inequality @xmath329 , ( c ) is due to the constraint in equation ( [ cons2 ] ) , and ( d ) follows from the fact that @xmath330 is equal to some joint distribution @xmath331 @xcite . since the chain of inequalities is true for every @xmath326 , we can choose the one that achieves @xmath4 , and then divide by @xmath1 to obtain the inequality in equation ( [ lowb3 ] ) in our theorem . to complete the proof of theorem [ thrdmax ] , we need to show that equality holds in the chain of inequalities above for some @xmath218 that satisfies the constraint .
if so , let us denote by @xmath332 the conditional pmf that achieves @xmath4 .
further , we denote by @xmath333 the corresponding causal conditioned pmf .
now , consider the following chain of equalities .
@xmath334}+\sum_{x^n}p(x^n)\log\gamma(x^n)\nonumber\\ & = -\lambda d+\sum_{x^n}p(x^n)\log\gamma(x^n),\nonumber\end{aligned}\ ] ] where ( a ) is due to a step in the algorithm given by ( [ step1 ] ) , and by the uniqueness of @xmath332 in the algorithm , as shown in lemma [ runique ] , and ( b ) follows the expression for @xmath218 given by @xmath335 therefore , we are left with verifying that the @xmath218 above satisfies the constraint : @xmath336 where ( a ) follows from equation ( [ step1 ] ) , and ( b ) is due to the causal conditioning chain rule .
hence , we showed that @xmath4 is the solution to the optimization problem given in equation ( [ lowb3 ] ) .
we also present an alternative proof for theorem [ thrdmax ] , this using the lagrange duality , as in @xcite , @xcite .
recall that @xmath4 is the result of @xmath337 where @xmath325 is defined by @xmath338 , subject to the following conditions : @xmath339 let us define the lagrangian as @xmath340 where @xmath341 for all @xmath342 .
differentiating the lagrangian , @xmath343 , over the variable @xmath326 , we obtain @xmath344 solving the equation @xmath345 in order to find the optimum value , yields the following expression @xmath346 where @xmath347 . multiplying both sides by @xmath348
we are left with the constraint @xmath349 where @xmath350 is induced by @xmath351 . from (
* chapter 5.1.3 ) we know that @xmath352 is a lower bound to @xmath4 . substituting the minimizer @xmath326 using equation ( [ minimizer ] ) , and the condition given by equation ( [ cond ] ) into @xmath353
, we obtain the lagrange dual function @xmath354{l l } $ -\lambda d+\sum_{x^n}p(x^n)\log\gamma'(x^n),$ & $ p(x^n)\gamma'(x^n)2^{-\lambda d(x^n,\hat{x}^n)}\leq p(x^n||\hat{x}^n)$\\ $ -\infty,$ & otherwise.\end{tabular}\right.\label{analsm}\ ] ] by making the constraints explicit , and since the minimization problem is convex , we obtain the lagrange dual problem , i.e. , @xmath4 is the solution to @xmath355 subject to @xmath356 for the @xmath350 that is induced by @xmath351 , and @xmath326 is the optimal pmf .
we use the notation of an _ optimal _ pmf if it achieves the optimal value .
for example , the pmf @xmath326 that achieves the minimum of the directed information given the distortion constraint , is optimal .
we say that the pmf , @xmath350 is optimal , if it is induced by the optimal @xmath326 .
another example is the maximization problem in ( [ maxgp ] ) .
we say that @xmath357 are optimal if they achieve the maximum value .
therefore , @xmath350 is optimal as well if it satisfies equation ( [ cond ] ) .
now , we wish to substitute the constraint to @xmath358 for some @xmath217 .
first , note that we always achieve equality in ( [ condsome ] ) since we can increase the value of @xmath218 and thus increase the objective .
this , combined with the fact that for @xmath359 , @xmath360 must be zero , we have equality in ( [ cond ] ) as well ( if @xmath361 , then @xmath362 , and equation ( [ minimizer ] ) holds too ) .
now , let us assume that the maximum in ( [ maxgp ] ) with the constraint in ( [ condsome ] ) is achieved at a _ non - optimal _
@xmath217 , i.e. , one that is not achieved using the optimal @xmath357 .
thus , the value obtained in ( [ maxgp ] ) is larger then the value achieved by @xmath350 , i.e. , @xmath4 ( since the maximization includes @xmath350 ) .
however , from the lagrange duality it should be a lower bound to @xmath4 , thus contradicting the fact that the maximum is achieved at a non - optimal @xmath217 .
note , that we can construct the optimal pmf @xmath326 from the solution to the maximization problem presented here .
consider the parameters @xmath363 that achieve ( [ maxgp ] ) , and calculate @xmath350 according to equation ( [ cond ] ) .
the calculation of @xmath326 is done recursively on @xmath364 . for @xmath365 , calculate @xmath366 using @xmath367 further , calculate @xmath368 using @xmath369 now , once we have @xmath370 for every @xmath371 , calculate @xmath364 using @xmath372 \frac{\sum_{x_i}p(x^i)r(\hat{x}^i|x^i)}{p(x^{i-1})r(\hat{x}^{i-1}|x^{i-1})},\nonumber\end{aligned}\ ] ] and then @xmath373 do so until @xmath374 , and we obtain our optimal @xmath326 .
another lemma we wish to provide is the connection between the curve of @xmath4 and the parameter @xmath221 .
this lemma is similar to the one given by berger in ( * ? ? ?
2.5.1 ) for the case of no feed - forward . [ rtag ]
consider the expression for @xmath4 given by @xmath375 where @xmath218 and @xmath221 are the variables that maximize ( [ maxgp ] ) .
we have seen that @xmath218 is of the form @xmath376 hence , the slope at distortion @xmath23 is @xmath377 . the proof is given simply by differentiating the expression for @xmath4 .
@xmath378\nonumber\\ & = -\frac{\lambda}{n}+\frac{1}{n}\left[-d+\sum_{x^n}\frac{p(x^n)}{\gamma(x^n)}\frac{d\gamma(x^n)}{d\lambda}\right]\frac{d\lambda}{dd}.\nonumber\end{aligned}\ ] ] now , consider the following expression @xmath379 using the @xmath218 given above , we have @xmath380 and thus @xmath381 .
@xmath382 p(x^n)q^*(\hat{x}^n||x^{n-1})2^{-\lambda d(x^n,\hat{x}^n)}\nonumber\\ & = \sum_{x^n}\frac{d\gamma(x^n)}{d\lambda}p(x^n)\sum_{\hat{x}^n}q^*(\hat{x}^n||x^{n-1})2^{-\lambda d(x^n,\hat{x}^n ) } -\sum_{x^n,\hat{x}^n}p(x^n)q^*(\hat{x}^n||x^{n-1})2^{-\lambda d(x^n,\hat{x}^n)}\gamma(x^n)d(x^n,\hat{x}^n)\nonumber\\ & = \sum_{x^n}\frac{d\gamma(x^n)}{d\lambda}\frac{p(x^n)}{\gamma(x^n)}-\sum_{x^n,\hat{x}^n}p(x^n)r^*(\hat{x}^n|x^n)d(x^n,\hat{x}^n)\nonumber\\ & = \sum_{x^n}\frac{d\gamma(x^n)}{d\lambda}\frac{p(x^n)}{\gamma(x^n)}-d\nonumber\\ & = 0.\nonumber\end{aligned}\ ] ] hence , we can conclude that @xmath383\frac{d\lambda}{dd}\nonumber\\ & = -\frac{\lambda}{n}.\nonumber\end{aligned}\ ] ]
in this appendix we prove the existence of a sequence of upper and lower bounds to @xmath4 , the rate distortion function with feed - forward .
these bounds correspond to an iteration in alg .
[ algs ] , and both converge to @xmath4 . to this end , we present and prove a few supplementary claims that assist in obtaining our main goal . theorem [ thrdmax ] provides an alternating form ( lagrange dual form ) of an optimization problem achieving @xmath4 , that is proved in app [ apprdmax ] . in lemma [ lemlow2 ]
, we show that in each iteration we can obtain measures that satisfy the constraint in theorem [ thrdmax ] to form a lower bound , and that the bound is tight and achieved as the upper bound converges .
we also provide a proof for the existence of a an upper bound in each iteration .
before we begin , we recall that a step in alg .
[ algs ] is defined by the following equality @xmath384 we shall use this equality throughout the proof . as mentioned , we use theorem [ thrdmax ] that provides us with the following alternating optimization problem .
@xmath385 where @xmath218 satisfies the inequality constraint @xmath386 for some causal conditioned probability @xmath217 .
we now show that in each iteration in alg .
[ algs ] , choosing @xmath218 appropriately forms a lower bound for @xmath4 .
[ lemlow2 ] in the @xmath272th iteration in alg .
[ algs ] , by letting @xmath387 and @xmath388 and defining @xmath389 the constraint in equation ( [ cons1 ] ) is satisfied , and forms a lower bound given by @xmath390 furthermore , this lower bound is tight , and is achieved as @xmath391 converges to @xmath4 , where @xmath391 is the upper bound .
let us fix the parameter @xmath392 as in ( [ gammak1 ] ) .
hence , @xmath393 where ( a ) follows from the definition of a step in alg .
[ algs ] and given above in equation ( [ step ] ) , and ( b ) follow the chain rule of causal conditioning , and @xmath394 is a causal conditioned pmf .
hence , combined with ( [ gammak3 ] ) , we obtain @xmath395 thus , we can use theorem [ thrdmax ] , and obtain a lower bound for @xmath4 , i.e. , @xmath396\nonumber\\ & = \frac{1}{n}\left[-\lambda d+\sum_{x^n}p(x^n)\log\gamma'^k_{x^n } -\sum_{x^n}p(x^n)\log\left(\max_{\hat{x}^n , x^{n-1}}c^k_{\hat{x}^n , x^{n-1}}\right)\right]\nonumber\\ & = \frac{1}{n } \left[-\lambda d+\sum_{x^n}p(x^n)\log\gamma'^k(x^n)-\log\left(\max_{\hat{x}^n , x^{n-1}}c^k_{\hat{x}^n , x^{n-1}}\right)\right].\label{lowb2}\end{aligned}\ ] ] to complete the proof of this lemma , we are left to show that as @xmath272 increases , i.e. , the upper bound converges to @xmath4 , the lower bound is tight . for that matter
, we note that the pmfs that achieve the optimum value @xmath397 are unique , as shown in lemma [ runique ] .
thus , it is clear that @xmath398 and @xmath399 placing equation ( [ gammak2 ] ) and ( [ ck2 ] ) in equation ( [ lowb2 ] ) , as shown in theorem [ thrdmax ] , achieves equality instead of the chain of inequalities given .
thus @xmath4 is , in fact , the solution to the optimization problem given in equation ( [ lowb1 ] ) , and we have demonstrated the existence of the lower bound [ lemlow3 ] in the @xmath272th iteration in alg . [ algs ] , the upper bound to the rate distortion is given by @xmath400 where @xmath401}$ ] .
note , that if @xmath402 produces a distortion @xmath23 , then @xmath403}-\sum_{x^n}p(x^n)\log\sum_{\hat{x}'^n}q^{k-1}(\hat{x}'^n||x^{n-1})2^{-\lambda d(x^n,\hat{x}'^n)}- \sum_{x^n,\hat{x}^n}p(x^n)r^k(\hat{x}^n|x^n)\log\frac{q^{k}(\hat{x}^n||x^{n-1})}{q^{k-1}(\hat{x}^n||x^{n-1})}\nonumber\\ & \stackrel{(b)}{=}-\lambda d_k+\sum_{x^n}p(x^n)\log\gamma^k(x^n)- \sum_{x^n,\hat{x}^n}p(x^n)r^k(\hat{x}^n|x^n)\log{c^k_{\hat{x}^n , x^{n-1}}},\label{lastineq}\end{aligned}\ ] ] where ( a ) follows from the definition of a step in alg .
[ algs ] and is given above in equation ( [ step ] ) , and ( b ) follows from the definition of @xmath404 .
hence , we have formed an upper bound to the rate distortion as in the lemma .
note that the only inequality is in the first line of the chain , and is due to the fact that @xmath405 .
however , upon convergence , this inequality is tight
. we can now conclude our main objective in this appendix .
_ proof of lemma [ lembounds ] _ proving this lemma requires us to present upper and lower bounds that converge to @xmath4 .
lemma [ lemlow2 ] provides us with a lower bound and its tightness , whereas lemma [ lemlow3 ] provides us with a tight upper bound as well , as required .
the markov source is presented in fig .
[ stockmod ] above
. we can describe the process @xmath406 using the equation @xmath407 where @xmath408 , @xmath409 .
this allows us to evaluate @xmath410 : @xmath411 where @xmath412 is the stationary distribution of the source .
now , to find the rate distortion of this model , we start with the converse @xmath413 where ( a ) follows from the fact that conditioning reduces entropy , and ( b ) follows the fact that @xmath414 and @xmath415 increases with @xmath23 for @xmath416 . where @xmath424 must hold for the following equation @xmath425 i.e. , @xmath426 note , that under this construction , the source @xmath11 is still markovian .
further , from fig . [ markex1 ] we can see that @xmath427 forms a markov chain , and @xmath428 .
thus , we obtain equality in ( a ) , ( b ) in the above chain of inequalities , and hence showed that @xmath429 by taking @xmath1 to infinity we obtain @xmath430 | in this paper we consider the rate distortion problem of discrete - time , ergodic , and stationary sources with feed forward at the receiver .
we derive a sequence of achievable and computable rates that converge to the feed - forward rate distortion .
we show that , for ergodic and stationary sources , the rate @xmath0 is achievable for any @xmath1 , where the minimization is taken over the transition conditioning probability @xmath2 such that @xmath3}\leq d$ ] .
the limit of @xmath4 exists and is the feed - forward rate distortion .
we follow gallager s proof where there is no feed - forward and , with appropriate modification , obtain our result .
we provide an algorithm for calculating @xmath4 using the alternating minimization procedure , and present several numerical examples .
we also present a dual form for the optimization of @xmath4 , and transform it into a geometric programming problem .
alternating minimization procedure , blahut - arimoto algorithm , causal conditioning , concatenating code trees , directed information , ergodic and stationary sources , geometric programming , ergodic modes , rate distortion with feed - forward .
introduction[secintro ] the rate distortion function for memoryless sources is well known and was given by shannon in his seminal work@xcite .
shannon@xcite showed that the rate distortion function is the minimum of mutual information between the source @xmath5 and the reconstruction @xmath6 , where the minimization is over transition probabilities @xmath7 such that the distortion constraint is satisfied , i.e. , @xmath8}\leq d$ ] . in the case where the source is stationary and ergodic
, gallager@xcite showed that the rate distortion is the limit of the following sequence of rates .
each member of the sequence is the @xmath1th order rate distortion function , which is the solution of the following minimization problem @xmath9 the minimization is over all conditional probabilities @xmath2 such that the distortion constraint is satisfied , i.e. , @xmath3}\leq d$ ] .
gallager showed that the limit of the sequence @xmath10 exists and is equal to the infimum of the sequence .
the problem of source coding with feed - forward was introduced by weissman and merhav@xcite and by venataramanan and pradhan@xcite , and is depicted in fig .
[ rdprob ] . [ ] [ ] [ 1]@xmath11[][][1]decoder[][][1]@xmath12 [ ] [ ] [ 1]@xmath13 [ ] [ ] [ 1]delay @xmath14[][][1]encoder[][][1]delay @xmath14[][][0.8]@xmath15 , and needs to reconstruct the source within the constraint @xmath3}\leq d$].,width=377 ] weissman and merhav@xcite named the problem competitive predictions . in their work , they defined a set of functions that predict the following @xmath16 given the previous @xmath17 . after defining the _ loss function _ between @xmath16 and the prediction , the objective was to minimizing the expected loss over all sets of predictors of size @xmath18 . an important result in @xcite
is that in the case where the innovation process @xmath19 is i.i.d .
the distortion - rate with feed - forward function is the same as the distortion - rate function of @xmath20 , where there is no feed - forward .
in particular , if @xmath16 is an i.i.d . process , then @xmath21 and thus the distortion - rate with feed - forward for the source @xmath16 is the same as if there is no feed - forward .
venkataramana and pradhan@xcite gave an explicit definition of the rate distortion feed - forward for an arbitrary normalized distortion function and a general source .
their goal was to provide the rate @xmath22 of a source given a distortion @xmath23 using causal conditioning and directed information . the source of information
is modeled as the process @xmath24 and is encoded in blocks of length @xmath1 into a message @xmath25 . the message @xmath26 ( after @xmath1 time units )
is sent to the decoder that has to reconstruct the process @xmath27 using the message @xmath26 and causal information of the source with some delay @xmath14 as in fig .
[ rdprob ] . for that purpose , venkataramanan and pradhan@xcite defined the measures @xmath28 and @xmath29 the limsup in probability of a sequence of random variables @xmath27 is defined as the smallest extended real number @xmath30 such that @xmath31 , @xmath32=0,\nonumber\end{aligned}\ ] ] and the liminf in probability is the largest extended real number @xmath33 such that @xmath31 , @xmath34=0.\nonumber\end{aligned}\ ] ] the main result in @xcite is that for a general source @xmath27 and distortion @xmath23 , the rate distortion with feed - forward @xmath35 is given by @xmath36 where the infimum is evaluated over the set @xmath37 of probabilities @xmath38 that satisfy the distortion constraint .
moreover , if @xmath39 venkataramana and pradhan showed in @xcite , that @xmath40 the work of venkataramanan and pradhan has made a significant contribution since it gives a multi - letter characteristic for the rate distortion function with feed - forward . in @xcite
, they evaluated these formulas for a stock - market example and provided an analytical expression for the rate distortion function .
however , these types of formulas are still very hard to evaluate for the general case . in this paper
we show that assuming ergodicity and stationarity of the source , the rate distortion function with feed - forward and delay @xmath41 is upper bounded by @xmath4 , where @xmath42}\leq d } i(\hat{x}^n\rightarrow x^n).\label{rlim}\end{aligned}\ ] ] we further show that the limit of the sequence @xmath43 exists , is equal to @xmath44 , and is the rate distortion feed - forward function @xmath35 .
these expressions for @xmath4 are computable using a blahut - arimoto - type algorithm or using geometric programming , as demonstrated here . in most models with causal constraints , such as feedback channels or feed - forward rate distortion , the causal conditioning probability , as well as the directed information characterizes the fundamental limits . in order to address these models ,
the causal conditioning probability was introduced by massey@xcite and kramer@xcite and is defined as @xmath45 the difference between regular and causal conditioning is that in causal conditioning the dependence of @xmath46 on future @xmath47 is not taken into account .
following the causal conditioning probability , massey @xcite ( who was inspired by marko s work @xcite on bidirectional communication ) introduced the directed information , defined as @xmath48 the directed information was used by tatikonda and mitter@xcite , permuter , weissman , and goldsmith @xcite , and kim @xcite to characterize the point - to - point channel capacity with feedback .
it is shown that the capacity of such channels is characterized by the maximization of the directed information over the input probability @xmath49 . in a previous paper@xcite , we used these results and obtained bounds to estimate the feedback channel capacity using a blahut - arimoto - type algorithm ( baa ) for finding the global optimum of the directed information .
the main contribution of this work lies in extending the achievability proof given by gallager in @xcite to the case where feed - forward with delay @xmath41 exists .
the extension is done by using the causal conditioning distribution , @xmath50 , rather than the regular reconstruction distribution @xmath51 , in order to construct the codebook .
the proof given is for @xmath41 , but can be extended straightforwardly to any delay @xmath52 .
the difficulty in this modification is that while in @xcite the codebook was an ensemble of sequences ( code words ) from the reconstruction alphabet using @xmath51 , our codebook is an ensemble of code trees using @xmath50 .
this induced a major problem while showing that the probability of error is small , as discussed in section [ secachieve ] .
these difficulties were overcome by appropriate modification to gallager s proofs .
another contribution of this paper is the development of two optimization methods for obtaining @xmath4 ; a ba - type algorithm and a geometric programming(gp ) form .
the gp form is given as a maximization problem , which can be solved using standard convex optimization methods .
further , this maximization problem gives us a lower bound to the rate distortion with feed - forward , which helps us decide when to terminate the algorithm .
the remainder of the paper is organized as follows . in section [ secprobres ]
we describe the problem model , provide the operational definition of the rate distortion function with feed - forward , and state our main theorems . in section [ secachieve ] we show that @xmath4 is an achievable rate for all @xmath1 and any distortion @xmath23 , and in section [ secoperational ] we show that the limit of @xmath4 exists and is equal to the operational rate distortion function . in section [ secalggp ]
we present an alternative optimization problem for @xmath4 in a standard geometric programming form that can be solved numerically using convex optimization tools . in section
[ secalgs ] we give a description of the baa for calculating @xmath4 and present the algorithm s complexity and the memory required , and in section [ secder ] we derive the baa and prove its convergence to the optimum value .
numerical examples are given in section [ secex ] to illustrate the performance of the suggested algorithms .
problem statement and main results[secprobres ] in this section we present notation , describe the problem model and summarize the main results of the paper .
we first state the definitions of a few quantities that we use in our coding theorems .
we denote by @xmath53 the vector @xmath54 .
usually we use the notation @xmath55 for short .
further , when writing a probability mass function ( pmf ) we simply write @xmath56 .
an alphabet of any type is denoted by a calligraphic letter @xmath57 , and its size is denoted by @xmath58 . in the rate distortion problem with feed - forward of delay @xmath41 , as shown in fig .
[ rdprob ] , we consider a general discrete , stationary , and ergodic source @xmath27 , with the @xmath1th order probability distribution @xmath49 , alphabet @xmath57 and reconstruction alphabet @xmath59 . the normalized bounded distortion measure is defined as @xmath60 on pairs of sequences . a @xmath61 source code with feed - forward of block length @xmath1 and rate @xmath22 consists of an encoder mapping @xmath62 ,
@xmath63 and a sequence of decoder mappings @xmath64 , @xmath65 the encoder maps a sequence @xmath66 to an index in @xmath67 . at time @xmath68 , the decoder has the message that was sent and causal information of the source , @xmath69 , and reconstructs the @xmath68th symbol sent , @xmath46 .
a rate distortion with feed - forward pair @xmath70 is achievable if there exists a sequence of @xmath61-rate distortion codes with @xmath71}\leq d.\nonumber\end{aligned}\ ] ] the rate distortion with feed - forward function @xmath35 is the infimum of rates @xmath22 such that @xmath70 is achievable . in this paper , we define the mathematical expression for the rate distortion function as the following limit @xmath72 where @xmath4 is the @xmath1th order rate distortion function given by @xmath42}\leq d}i(\hat{x}^n\to x^n).\nonumber\end{aligned}\ ] ] we show that the limit in ( [ limoper ] ) exists , @xmath4 is achievable and upper bounds @xmath73 for all @xmath1 .
further , we show that the rate distortion feed - forward function , @xmath35 , is equal to @xmath73 .
we also provide two ways to calculate numerically the value @xmath4 ; using a ba - type algorithm and a geometric programming form .
we now state our main theorems .
[ thrd ] for a discrete , stationary , ergodic source , and for any @xmath23 , any @xmath1 and delay @xmath41 , @xmath4 is an achievable rate .
[ thoperat ] for any distortion @xmath23 , the operational rate distortion function @xmath35 is equal to the mathematical expression , @xmath73 , where @xmath73 is given by ( [ limoper ] ) .
[ thrdmaxgp ] the @xmath1th order rate distortion function @xmath4 can be written in a geometric programming standard form as the following maximization problem @xmath74 subject to the constraints : @xmath75 [ ths ] for a fixed source distribution @xmath49 , there exists an alternating minimization procedure in order to compute @xmath42}\leq d}i(\hat{x}^n\rightarrow x^n).\label{rd1}\end{aligned}\ ] ] proofs to theorem [ thrd ] and [ thoperat ] are given in section [ secachieve ] and section [ secoperational ] , respectively .
the proof for theorem [ thrdmaxgp ] is in section [ secalggp ] , the algorithm in theorem [ ths ] is described in section [ secalgs ] and proved in section [ secder ] .
achievability proof ( theorem [ thrd]).[secachieve ] in this section we show that if the source is stationary and ergodic , then @xmath4 as given in ( [ rd1 ] ) is achievable for any @xmath1 . in order to do so
, we first assume that the source is ergodic in blocks of length @xmath1 , and show achievability . a source that is
ergodic in blocks is one that , by looking at each @xmath1 letters as a single letter from a super alphabet , we obtain an ergodic super source ( presented in ( * ? ?
* chapter 9.8 ) ) .
then , for the general ergodic sources , we follow a claim given in @xcite about ergodic modes , as explained further on .
the distortion is assumed to be normalized , finite , and of the form @xmath76 for some @xmath77 .
an example for such a distortion can be found in @xcite and in section [ secex ] , in an example called the stock - market . [ thrdn ] consider a discrete stationary source that is ergodic in blocks of length @xmath1 . for any distortion @xmath23 such that @xmath78 and @xmath79 , and for any @xmath80 sufficiently large , there exists a codebook of trees @xmath81 of length @xmath80 with @xmath82 code trees for which the average distortion per letter satisfies @xmath83}\leq d+\delta$ ] .
let @xmath2 be the transition probability that achieves the minimum @xmath4 and let @xmath84 be the causal conditioning probability that corresponds to @xmath85 .
* for any @xmath80 , consider the ensemble of codes @xmath81 with @xmath86 code trees of length @xmath80 , where each code tree @xmath87 is a concatenation of @xmath88 sub - code trees of length @xmath1 . each sub - code tree is generated independently according to @xmath84 as in fig .
[ codetree ] .
+ [ ] [ ] [ 1]@xmath89 [ ] [ ] [ 1]@xmath90 [ ] [ ] [ 1]@xmath91 [ ] [ ] [ 1]@xmath92 [ ] [ ] [ 1]@xmath93 [ ] [ ] [ 1]@xmath94 [ ] [ ] [ 1]@xmath95 [ ] [ ] [ 1]@xmath96 [ ] [ ] [ 1]@xmath96 [ ] [ ] [ 1]@xmath97 [ ] [ ] [ 1]@xmath98 [ ] [ ] [ 1]code tree 1[][][1]code tree 2 + .
the upper branches are for @xmath99 , and the lower branches are for @xmath100 .
] * the encoder assigns a code tree @xmath87 for every @xmath101 such that @xmath102 is minimal .
the sequence @xmath103 is determined by walking on tree @xmath104 , and following the branch @xmath105 . * at time @xmath68 , the decoder possesses the index of the tree @xmath104 and causal information of the source @xmath69 , and returns the symbol @xmath106 that it produces .
let us define a test channel as the conditional probability @xmath107 and the causal conditional probability @xmath108 where the distribution is according to @xmath109 moreover , we define for every code tree @xmath104 of length @xmath80 the measure @xmath110 where @xmath111 .
note that @xmath112 is not the directed information between the sequences @xmath113 , but simply a measure between a source sequence @xmath101 and the output , @xmath114 of the test channel @xmath115 , as defined in ( [ testchan ] ) .
let @xmath116 be the set of all code trees of length @xmath80 , and consider the following set , @xmath117 and let @xmath118 be the probability of the set @xmath119 on the test channel ensemble .
let us use the notation @xmath120 where @xmath121 is the ensemble of code trees as described in the coding scheme .
now , let @xmath122 be the probability over the ensemble of codes @xmath121 and source sequences such that the distortion exceeds @xmath123 .
we wish to give an upper bound to the probability @xmath122 ; for this we use the following lemma . for a given source @xmath124 and test channel
, we have the following inequality @xmath125 where the set @xmath126 is described in ( [ testa ] ) . _ proof . _
we first write @xmath127 as @xmath128 for every @xmath101 , let us define the set @xmath129 as the set of all code trees @xmath130 for which @xmath131 , @xmath132 we observe that @xmath133 for a given @xmath101 only if @xmath134 for every @xmath135 .
thus , @xmath133 only if @xmath136 for every @xmath137 .
since @xmath104 are independently chosen , @xmath138 where @xmath139 is the complement set of @xmath129 .
we note that the probability that tree @xmath104 being in @xmath139 depends only on the branch associated with @xmath140 . in other words , if a tree @xmath141 , then all other trees with the same branch associated with @xmath101 is in @xmath139 as well ; the same goes for @xmath129 .
hence , we can divide the set of all code trees @xmath142 into disjoint subsets @xmath143 that have the same branch associated with @xmath105 , i.e. , @xmath144 where @xmath145 is a walk on tree @xmath104 over the branch @xmath105 . clearly , the probability of each subset @xmath143 is @xmath146 since the left hand side is a summation of the probabilities of all trees with the same branch associated with @xmath101 , and we are left with the probability of that one branch .
now , for every @xmath147 , and due to the definition of @xmath139 , we have @xmath148 therefore , @xmath149 and we obtain that @xmath150 where ( a ) follows the inequality in equation ( [ ineqach ] ) . using the inequality @xmath151 , and taking @xmath152 , @xmath153
, we find @xmath154 by taking a sum over @xmath101 we remain with @xmath155 note , that @xmath156 where ( a ) follows the fact that if @xmath157 , then @xmath114 is determined by the tree @xmath104 and the branch @xmath101 . now ,
continuing from equation ( [ sumprob ] ) , we obtain @xmath158 we now use the result in ( [ per ] ) in order to complete the proof of the theorem .
furthermore , we can see that the average distortion of the code satisfies @xmath159}(\leq ( d+\delta/2)+p_c\left(d(x^l,\hat{x}^l({\cal{t}}_c , x^{l-1}))>l(d+\delta/2)\right)\cdot\sup_{x^l,\hat{x}^l}{d(x^l,\hat{x}^l)}.\nonumber\end{aligned}\ ] ] this arises , as in ( * ? ? ?
* th .
9.3.1 ) , from upper bounding the distortion by @xmath160 when the @xmath161 , and by @xmath162 otherwise . by choosing @xmath86 , the last term in ( [ per ] )
goes to zero with increasing @xmath80 .
furthermore , the first term is bounded by @xmath163 note that @xmath164 as assumed , the source is ergodic in blocks of length @xmath1 .
furthermore , the test channel is defined to be memoryless for blocks of length @xmath1 , and hence the joint process is ergodic in blocks of length @xmath1 .
thus , with probability 1 , @xmath165}\nonumber\\ & = r_n(d).\nonumber\end{aligned}\ ] ] therefore , the probability of the first term in ( [ pabound ] ) goes to zero as @xmath80 goes to infinity , and the same goes to the second term due to the definition of the distortion .
in order to finish the proof , and due to the fact that @xmath166 goes to zero with increasing @xmath80 and the fact that the distortion is finite , we can choose @xmath80 large enough such that @xmath167 in this case , we obtain @xmath168 , and hence the rate @xmath4 is achievable for sources that are ergodic in blocks of length @xmath1 . much like in gallager s proof for the case where there is no feed - forward , we note that not all ergodic sources are also ergodic in blocks , and we need to address these cases as well . for that purpose ,
we need ( * ? ? ?
* lemma 9.8.2 ) for ergodic sources .
we recall , that a discrete stationary source is ergodic if and only if every invariant set of sequences under a shift operator @xmath26 is of probability 1 or 0 . in (
* ? ? ?
* chapter 9.8 ) , the author looks at the operator @xmath169 , i.e. , a shift of @xmath1 places , and considers an invariant set @xmath170 , @xmath171 , with respect to @xmath169 . in lemma 9.8.2 in @xcite , it is stated that one can separate the source @xmath172 to @xmath173 invariant subsets @xmath174 , @xmath175 , with regard to @xmath169 , such that @xmath173 divides @xmath1 and the sets @xmath176 are disjoint except , perhaps , an intersection of zero probability .
these subsets are called _ ergodic modes _ , due to the fact that each invariant subset of them under the operator @xmath169 is of probability 0 or @xmath177 . in other words ,
conditional on an ergodic mode @xmath178 each invariant subset of it with respect to @xmath169 , is of probability 0 or 1 .
recall , that by definition , @xmath179 where the right - hand side is the average directed information between the source and reconstruction , determined according to @xmath2 that achieves @xmath4 .
let @xmath180 be the average directed information between a source sequence from the @xmath68th ergodic mode and the ensemble of codes , using the probability @xmath2 which achieves @xmath4 .
note that the directed information can be written as @xmath181 which is convex over the input probability @xmath49 .
thus , @xmath182 we observe that @xmath183 is an upper bound to the @xmath1th order rate distortion function conditional on the @xmath68th ergodic mode . from theorem [ thrdn ] , we know that there exists a codebook @xmath184 with @xmath185 code trees of length @xmath80 such that the average distortion constraint holds .
another observation is that if a codebook @xmath184 satisfies the distortion constraint , conditional on the ergodic mode @xmath178 , then it has the same effect conditional on the ergodic mode @xmath186 . in other words
, we can encode not only a source sequence from @xmath187 with @xmath188 , but also a shift of the a source sequence in @xmath189 with @xmath190 .
we use these observations while constructing the codebook
. we can now prove theorem [ thrd ] , i.e. , the achievability of @xmath4 , where the source is ergodic and stationary .
an equivalent version of theorem [ thrd ] is the following : let @xmath4 be the @xmath1th order rate distortion function for a discrete , stationary , and ergodic source . for any @xmath23 such that @xmath78 , and @xmath79 , and any @xmath80 sufficiently large , there exists a codebook of trees @xmath81 of length @xmath80 with @xmath82 code trees for which the average distortion per letter satisfies @xmath191}\leq d+\delta$ ] .
let @xmath2 be the transition probability that achieves @xmath4 and let @xmath84 be the causal conditioning probability that corresponds to @xmath85 .
* for any @xmath80 and any ergodic mode @xmath178 , @xmath192 , construct an ensemble of codes @xmath184 , with @xmath185 little code trees of length @xmath80 , where each little code tree is generated according to @xmath193 , as in fig .
[ codetree ] in theorem [ thrdn ] above .
now , for every @xmath192 , the @xmath68th codebook is an ensemble of big code trees , which are concatenation of @xmath173 little code trees , starting from one in @xmath184 , and followed by one from @xmath194 to one from @xmath195 , where the index is calculated modiolus @xmath173 . in the example of a big code tree in fig .
[ ncodetree ] we see additional letters at the end of each little code tree , i.e. , in positions @xmath196 , that are fixed .
the purpose of the fixed letters is to shift the sequence and encode it with a codetree from the sequential codebook .
note , that the overall length of a code tree sums up to @xmath197 .
+ [ ] [ ] [ 1]codetree from @xmath184 [ ] [ ] [ 1]codetree from @xmath194 [ ] [ ] [ 1]codetree from @xmath198 [ ] [ ] [ 1]fixed letters [ ] [ ] [ 1]code tree 1[][][1]code tree 2 + th codebook , @xmath199 , @xmath200 . ]
* for every @xmath68 , the encoder assigns for every source sequence @xmath201 a code tree @xmath202 from the @xmath68th codebook , such that @xmath203 is minimal .
the sequence @xmath204 is determined by walking on tree @xmath202 , and following the branch @xmath205 . *
the decoder receives a tree @xmath202 and causal information of @xmath206 and returns the sequence @xmath207 that it produces . since the distortion constraint for every ergodic mode
is satisfied due to theorem [ thrdn ] , the overall distortion is satisfied as well .
the additional fixed letters are of unknown distortion , but due to the face that the distortion is bounded , their contribution is negligible for large @xmath80 .
moreover , note that for every @xmath68 , the @xmath68th codebook is of the same size .
thus , the overall size of the codebook is @xmath208 recall that @xmath197 , and by letting @xmath209 we conclude that @xmath4 is an achievable rate for the general ergodic source , as required .
proof that @xmath210 ( theorem [ thoperat]).[secoperational ] in this section we show that the operational description of the rate distortion with feed - forward is equal to the mathematical one given in ( [ rdi ] ) .
this will be done first by showing that the mathematical expression @xmath73 is achievable , and then by showing that it is a lower bound to the rate distortion function .
we recall that @xmath211}\leq d}i(\hat{x}^n\to x^n).\label{rdi}\end{aligned}\ ] ] to show that @xmath73 is achievable we first need to show that the limit of the sequence @xmath43 exists . for this purpose ,
we use the following lemma .
[ lemsubadd ] the sequence @xmath4 , @xmath42}\leq d}i(\hat{x}^n\rightarrow x^n),\nonumber\end{aligned}\ ] ] is a sub - additive sequence , and thus @xmath212 note , that a sequence @xmath213 is called sub - additive if for all @xmath214 , @xmath215 the proof for lemma [ lemsubadd ] is given in app .
[ appsubadd ] .
we now state a lemma for the achievability of @xmath73 .
[ lemach ] the mathematical expression for the rate distortion feed - forward @xmath73 is achievable , and thus upper bounds @xmath35 .
we showed in theorem [ thrd ] that for any @xmath1 , @xmath4 is achievable .
further , in lemma [ lemsubadd ] we show that the limit exists and equal to the infimum , and hence is achievable too .
therefore , we conclude that the mathematical expression @xmath73 is achievable , and forms an upper bound to the operational description @xmath35 . to show that @xmath73 is a lower bound to the rate distortion function , we provide the following lemma [ lemconverse ] the mathematical expression @xmath73 is a lower bound to the operational rate distortion function . for the completeness of the paper , we provide the proof of lemma [ lemconverse ] , this in app .
[ appconv ] .
however , similar proof was presented by venkataramana and pradhan in @xcite , and their expressions involved limit in probability of the entropy and directed information as described in section [ secintro ] .
combining lemmas [ lemach ] , [ lemconverse ] provides us with the proof for our fundamental theorem , stated in section [ secprobres ] , i.e. , the operational rate distortion function @xmath35 is equal to the mathematical one , @xmath73 . geometric programming form to @xmath4 ( theorem [ thrdmaxgp])[secalggp ] in this section
we show that the @xmath1th order rate distortion function with feed - forward @xmath4 can be given as a maximization problem , written in a standard form of geometric programming . for this purpose we first state the following theorem .
[ thrdmax ] the @xmath1th order rate distortion function , @xmath4 , can be written as the following maximization problem @xmath216 where , for some causal conditioned probability @xmath217 , @xmath218 satisfies the inequality constraint @xmath219 in app .
[ apprdmax ] we provide two proofs for theorem [ thrdmax ] ; the first is similar to berger s proof in @xcite for the regular rate distortion function based on the inequality @xmath220 , and the second uses the lagrange duality as presented in @xcite and @xcite that transforms a minimization problem to a maximization one .. app .
[ apprdmax ] also includes the connection between the rate distortion function and the parameter @xmath221 , which states that the slope of @xmath4 in point @xmath23 is @xmath222 . considering the theorem above
, our interest now is to adjust the constraints in order to obtain a geometric programming form .
we note that the optimization problem in ( [ lowb4 ] ) does not change if we maximize over @xmath217 as well , and the constraint ( [ cons3 ] ) is no longer for some @xmath223 , i.e. , @xmath224 where @xmath225 satisfy the inequality constraint @xmath226 the above statement is true since , on the one hand , the maximization in ( [ lowb4 ] ) increases upon maximizing over another variable , @xmath217 , as in ( [ lowb5 ] ) ; on the other hand , the variable @xmath227 that achieves ( [ lowb5 ] ) satisfy the constraint ( [ cons3 ] ) in theorem [ thrdmax ] , and hence the maximization problem in ( [ lowb5 ] ) can not be greater than the one in ( [ lowb4 ] ) . to obtain a geometric programming standard form
we transform the constraint in ( [ cons4 ] ) , such that @xmath228 taking the @xmath229 of both sides , we obtain @xmath230 note that maximizing over @xmath217 is the same as maximizing over its products @xmath231 ( * ? ? ?
* lemma 3 ) .
therefore , we can conclude that the rate distortion with feed - forward @xmath4 can be given as a geometric programming maximization form , @xmath232 subject to @xmath75 hence , we obtain a standard form of geometrical programming .
this gp problem can be solved using standard convex optimization tools .
extension of the baa for rate distortion with feed - forward[secalgs ] in this section we describe an algorithm for calculating @xmath4 , where @xmath233}\leq d}i(\hat{x}^n\rightarrow x^n),\label{rlim3}\end{aligned}\ ] ] using the alternating minimization procedure .
this method was first used by blahut and arimoto @xcite , @xcite to obtain a numerical solution for the i.i.d . source rate distortion and for the memoryless channel capacity .
recently , in @xcite we extended this method for finding the global maximum of the following optimization problem- @xmath234 and we apply similar methods here .
before we describe the algorithm , let us denote by @xmath235 the pmfs that are participating in the minimization .
further , let us consider the double optimization problem given by @xmath236,\label{rlimdouble}\end{aligned}\ ] ] where @xmath237},\nonumber\end{aligned}\ ] ] and @xmath238 is the directed information that can be written as @xmath239 in section [ secder ] we show that the double optimization problem given in ( [ rlimdouble ] ) is equal to the one given in ( [ rlim3 ] ) . equations ( [ rlimdouble ] ) , ( [ directed4 ] ) allow us to apply the alternating minimization procedure
. description of the algorithm in algorithm [ algs ] we present the steps required to minimize the directed information where the input pmf @xmath49 is fixed .
* fix a value of @xmath240 that determines a point on the @xmath4 curve . + * start from a random causally conditioned point @xmath241 .
usually we start from a uniform distribution , i.e. , @xmath242 for every @xmath243 . + * set @xmath244 . + * compute @xmath245 using the formula @xmath246 * calculate the joint probability @xmath247 , and deduce the causal conditioned pmf @xmath248 as in ( [ causalpr ] ) . + * calculate the parameter @xmath249 * calculate @xmath250 * if @xmath251 , set @xmath252 , and return to ( d ) . +
* the rate distortion function , with distortion @xmath253 , is @xmath254 the parameter @xmath221 is used in the lagrangian with which we optimize the directed information .
the value of @xmath255 and hence @xmath256 depends on @xmath221 ; thus choosing @xmath221 appropriately sweeps out the @xmath256 curve .
the algorithm stops when @xmath257 . in app .
[ bounds ] we provide upper and lower bounds , used show that if @xmath257 , we ensure that @xmath258 .
now , let us present a special case and a few extensions for algorithm [ algs ] .
* _ regular baa , i.e. , the delay @xmath259_. for delay @xmath259 , the algorithm suggested here meets the original baa , where instead of step ( d ) we have @xmath260 and in step ( e ) , @xmath261 corresponds to the joint probability @xmath262 as well .
moreover , the expression for @xmath263 is reduced to @xmath264 and the termination of the algorithm in step ( g ) is defined by @xmath265 as in the regular blahut - arimoto algorithm @xcite . *
_ function of the feed - forward with general delay @xmath14_. we present a generalization of the algorithm , where the feed - forward is a deterministic function of the source with some delay @xmath14 , @xmath266 . in that case ,
step ( d ) is replaced by @xmath267 and in step ( e ) we have @xmath268 where we calculate @xmath269 from the joint distribution @xmath270 .
the algorithm is terminated in the same way , where @xmath271 complexity and memory needed computation complexity and memory needed for the algorithm above is presented in table [ complexmemory2 ] .
.memory and operations needed extended baa for source coding with feed - forward . [
cols="^,^,^",options="header " , ] [ tt3 ] it was shown in @xcite that the rate - distortion function of a general markov - chain source with @xmath272 states , is given by @xmath273 where @xmath274 $ ] is the stationary distribution of the markov chain , and @xmath275 . in our special case
we have @xmath276 , i.e. , @xmath277 states for the markov chain , and transition probabilities @xmath278 as illustrated in fig .
[ stockmod ] .
the stationary distribution of such a source is @xmath279 $ ] , and we are left with @xmath280 since the rate can not be less than zero , and is a descending function of the distortion , the rate - distortion function is as above when @xmath281 , i.e. , when @xmath282 , and thus we obtain @xmath283{l l } $ 0.6(h_b(0.2)-h_b(\frac{d}{0.6})),$ & $ d\leq0.12$\\$0,$ & otherwise.\end{tabular}\right.\label{analsm}\ ] ] in fig .
[ smfig](a ) we present the graphs of @xmath4 for @xmath284 up to @xmath285 with the distortion described here and where @xmath286 has the stationary distribution @xmath287 $ ] .
we can see that @xmath4 decreases as @xmath1 increases as expected and converges to the analytical calculation . in fig .
[ smfig ] ( b ) we present the directed information rate estimator only for @xmath285 , where the vector of the distortion is interpolated , i.e. , @xmath288 .
we can see that this estimator is much more accurate than the one in fig .
[ smfig ] ( a ) .
[ ] [ ] [ 0.8]@xmath35 [ ] [ ] [ 0.8]@xmath23 [ ] [ ] [ 0.8]@xmath1 [ ] [ ] [ 0.8 ] [ ] [ ] [ 0.8]@xmath35 [ ] [ ] [ 0.8]@xmath23 \(a ) graph of @xmath4 ; the arrow marks the way @xmath4 responds to @xmath1 increasing .
the dashed line is the analytical calculation .
+ ( b ) graph of @xmath288 .
the circles represent the performance of alg .
[ algs ] .
[ smfig ] the effects of the delay on @xmath4 in this example we use the markov source ( fig .
[ stockmod ] ) example with a single letter distortion .
we run alg .
[ algs ] with delays @xmath289 and block length @xmath290 , where @xmath286 has the stationary distribution .
we expect the rate distortion function to increase with the delay @xmath14 .
this is expected because as the delay @xmath14 increases the value of the directed information increases as well . due to the fact that for @xmath291 all graphs are close together , we present @xmath4 only for @xmath292 , and the results are shown in fig .
[ markdelay ] .
[ ] [ ] [ 0.8]@xmath35 [ ] [ ] [ 0.8]@xmath23 [ ] [ ] [ 0.8]@xmath293 for a markov source as a function of the delay.,width=226 ] conclusions in this paper we considered the rate distortion problem of discrete - time , ergodic , and stationary sources with feed forward at the receiver .
we first derived a sequence of achievable rates , @xmath294 , that converge to the feed - forward rate distortion . by showing that the sequence is sub - additive
, we proved that the limit of @xmath4 exists and thus equals to the feed - forward rate distortion .
we provided an algorithm for calculating @xmath4 using the alternating minimization procedure , and also presented a dual form for the optimization of @xmath4 , and transformed it into a geometric programming maximization problem . |
narrow - line seyfert 1 galaxies are a particular class of active galactic nuclei ( agn ) that show : ( i ) small width of their broad optical emission lines , i.e. fwhm(h@xmath1kms@xmath2 @xcite , pointing to a low black hole ( bh ) mass for these systems ( @xmath3m@xmath4 ) , ( ii ) super - strong feii emission complexes , ( iii ) rapid x - ray variability , hinting that the broad - line region ( blr ) and accretion disk are directly visible , ( iv ) near - eddington accretion rates with ratios @xmath5 between 0.1 and 1 @xcite , ( v ) super - soft x - ray spectra , and ( vi ) a multitude of intriguing multi - wavelength properties ( see review by * ? ? ?
there exists also an interesting small fraction of their population being radio - loud , featuring relativistic jets , and detected at @xmath0 rays by _ fermi _
( @xmath0nls1s ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these few sources , the radio- and @xmath0-ray - loud population , are exceptional because they show blazar - like observational attributes such as flat radio spectra , high brightness temperatures ( t@xmath6k ) , doppler boosting , @xmath0-ray emission and in few cases , one - sided jets . however ,
possessing non - blazar physical properties , especially almost two orders of magnitude lower bh masses and high accretion rates . in this context
, they constitute a source of new insights , able to give us clues on the formation of extragalactic jets and their evolution under conditions of high accretion rate and in a regime that is not probed by classical blazars . as possible explanations for their appearance ,
different scenarios have been proposed .
do they represent a young agn population that rapidly grow their bhs ? and
how their orientation , with respect to the observer s line of sight , enters this challenging equation ? in the following , we present our radio monitoring of @xmath0nls1s using primarily the 100-m telescope at effelsberg and the iram 30-m telescope , while focusing on latest results .
furthermore , we focus on two @xmath0nls1s which have been studied in detail : ( i ) rxj2314.9 + 2243 with dedicated multi - wavelength monitoring , and ( ii ) 1h0323 + 342 with the help of high - resolution very - long - baseline interferometry ( vlbi ) .
our monthly monitoring of the jet emission from @xmath0-ray - emitting nls1 galaxies constitutes the most comprehensive such program at cm and short mm wavelengths .
observations with the effelsberg 100-m telescope cover eight bands between 2.64 and 43.05ghz . until 2014 ,
the iram 30-m telescope provided coverage at 86.24 and 142.33ghz . the data collected in more than five years is the longest data set with the widest frequency coverage available for these objects ( see fig .
[ fig1 ] ) .
it includes also radio polarisation monitoring at 2.64 , 4.85 , 8.35 , and 10.45ghz , with measurements of its linear and circular components and evpas .
the observed sample comprises the following sources : 1h0323 + 342 ( j0324 + 3410 ) , sbs0846 + 513 ( j0849 + 5108 ) , pmnj0948 + 0022 ( j0948 + 0022 ) , and pks1502 + 036 ( j1505 + 0326 ) .
recently , three more @xmath0nls1s were included , namely fbqsj1644 + 2619 , sdssj1222 + 0413 , and b31441 + 476 .
nls1s between 2.6 and 42ghz , until 2015 january .
where available , data at 86 and 142.33ghz are also shown.,title="fig:",scaledwidth=49.0% ] nls1s between 2.6 and 42ghz , until 2015 january .
where available , data at 86 and 142.33ghz are also shown.,title="fig:",scaledwidth=49.0% ] + nls1s between 2.6 and 42ghz , until 2015 january .
where available , data at 86 and 142.33ghz are also shown.,title="fig:",scaledwidth=49.0% ] nls1s between 2.6 and 42ghz , until 2015 january . where available , data at 86 and 142.33ghz are also shown.,title="fig:",scaledwidth=49.0% ] all sources show the typical blazar behaviour although with lower flux densities .
@xmath0nls1s are variable and flare often , but the events are characterized by smaller amplitude and shorter duration .
however , more energetic outbursts can be seen from time to time which dominate a source s light curve .
their spectra exhibit intense evolution with time , and steep , flat or even inverted spectra can be seen at different epochs .
this evolution occurs faster than in blazars and can be ascribed to evolving shocks .
the estimates of their jet powers are comparable with the output of the least energetic blazars .
furthermore , their moderate t@xmath7 and associated doppler factors ( @xmath8 ) , point to mildly relativistic jets as the driver of their phenomenology .
most sources do not display detectable polarisation .
however , 1h0323 + 342 is @xmath95.5% linearly polarized , higher than what is typically observed in blazars ( e.g. @xmath93.55% * ? ? ?
its evpa , at @xmath10 , lies perpendicular to the jet orientation , suggesting a magnetic field orientation almost parallel to the jet .
pks1502 + 036 appears to be polarized at 8.35 ghz ( @xmath923% with its evpa at @xmath11 ) during flaring periods . in the r band , @xmath0nls1s
have been monitored with robopol since 2013 ( e.g. * ? ? ?
their mean linear polarisation ranges from less than 1% to up to 20% .
this is a mildly radio - loud ( r = 1020 ) nls1 galaxy at a redshift z = 0.1692 with a bh mass of 8@xmath1210@xmath13m@xmath4 and a ratio @xmath14 @xcite .
what sets this source apart from the rest of the nls1 population is its tentative detection at @xmath0 rays by _ fermi _ ( l. foschini , priv .
comm . ) and its steep radio spectrum .
the ongoing radio observations verify the steep spectrum with flux densities @xmath15mjy and @xmath16mjy on two occasions in 2014 october . to date , all @xmath0-ray detected nls1s have been associated with flat spectrum sources and rxj2314.9 + 2243 could be the first steep radio spectrum nls1 to emit @xmath0 rays .
address the emission processes shaping the spectral energy distribution ( sed ) of rxj2314.9 + 2243 .
the source is a luminous infrared emitter with an sed showing a broad emission hump between ir and uv bands , that steepens in the uv .
an unusual feature of the source , compared to other nls1s with steep x - ray spectra , is the flatness of its x - ray spectrum , hinting at jet activity . in the uv on the other hand ,
the spectrum is very steep but there is no evidence of optical extinction beyond what is expected from our galaxy , as would be expected by the dusty environment of luminous infrared galaxies ( lirgs ) .
therefore the ir to uv emission is most likely of non - thermal ( i.e. synchrotron ) origin arising from the operating jet .
another striking feature of rxj2314.9 + 2243 is its very broad o[iii ] 5007 emission line component .
this shows an unusually high kinematical blueshift of 1260 kms@xmath2 that could be due to the presence of an outflow powered by a radio jet feedback , in a face - on orientation with respect to our line of sight .
1h0323 + 342 is the most nearby radio - loud ( r @xmath17 ) @xmath0-ray emitting nls1 at z @xmath18 @xcite .
it exhibits a high eddington luminosity ratio ( @xmath19 ) with a low bh mass of @xmath20m@xmath4 confirmed by several independent estimates @xcite .
1h0323 + 342 is highly variable at radio bands ( cm to mm ; see fig .
[ fig1 ] ) and is also a special case because of the morphology of its host galaxy which appears ring - like or as a one - armed spiral .
we studied the parsec - scale structure and kinematics of the source employing high - resolution vlbi images at 15ghz obtained in eight epochs from 2010 october until 2013 july .
additionally , with our single - dish observations we infer the t@xmath7 and @xmath21 for the source .
this combination provides a good estimate for the viewing angle towards 1h0323 + 342 ( see * ? ? ?
* for first results ) .
1h0323 + 342 shows a core
jet morphology characterized by the prominent core , a straight jet , and a total absence of emission for the receding side of the jet ; i.e. no counter - jet is detected .
a stationary feature is seen close to the core while six components move along the relativistic jet .
five jet components show superluminal motion with velocities from @xmath22c up to @xmath23c @xcite .
the source shows fast variability seen with single - dish and vlbi monitoring alike .
the flaring events are of low to mild amplitude , but more rapid compared to the typical long - term behavior of blazars . 1h0323 + 342 shows high brightness temperatures . while report a doppler factor of 3.6 at 14.6ghz , and a higher one @xmath24 at 2.64ghz , we deduce a more extreme value of t@xmath25k and an associated doppler factor of 5.2 , from the most rapid and largest flux density variation . combining the knowledge on the source kinematics ( i.e. the speeds of jet components ) with the doppler factors inferred from the observed variability , allowed us , under the assumption of causality , to constrain the viewing angle towards the source .
the range of plausible values is @xmath26@xmath27 @xcite .
the smaller value of this range is consistent with viewing angles typical of blazars , but nevertheless a larger @xmath28 with the respect to the observer s line of sight can not be excluded .
the remarkable characteristics of nls1 galaxies separate them as prominent class of agn and made them , in recent years , the focal point of intensive research .
these systems accrete matter at a rate near the eddington limit and rapidly grow the low - mass black holes driving them .
the study of the small population of radio - loud and gamma - ray detected nls1s probes a parameter space not constrained by classical blazars and can provide a deeper understanding on the formation and evolution of powerful radio jets and the interplay with their complex environments . | narrow - line seyfert 1 ( nls1 ) galaxies provide us with unique insights into the drivers of agn activity under extreme conditions . given their low black hole ( bh ) masses and near - eddington accretion rates
, they represent a class of galaxies with rapidly growing supermassive bhs in the local universe . here , we present the results from our multi - frequency radio monitoring of a sample of @xmath0-ray loud nls1 galaxies ( @xmath0nls1s ) , including systems discovered only recently , and featuring both the nearest and the most distant @xmath0nls1s known to date .
we also present high - resolution radio imaging of 1h 0323 + 342 , which is remarkable for its spiral or ring - like host .
finally , we present new radio data of the candidate @xmath0-emitting nls1 galaxy rx j2314.9 + 2243 , characterized by a very steep radio spectrum , unlike other @xmath0nls1s . |
semiconductor cavities have attracted considerable attention in recent years due to their ability to store light for a given amount of time in a small volume @xcite .
this key issue of cavities stimulated a large amount of experiments for increasing the nonlinear interaction of photons and to understand and exploit cavity quantum electrodynamics ( cqed ) effects @xcite .
the dynamic manipulation of these systems , especially of combined cavity emitter systems is thereby of major interest @xcite .
all - optical switching of cavities gains momentum since it enables the dynamic control of the capture and release of photons on sub - picosecond timescales @xcite .
moreover , ultrafast change of the optical properties of a cavity prevails to frequency conversion through adiabatic @xcite and not - adiabatic processes @xcite .
the optical properties of cavities can be altered by changing the refractive index of the constituent material .
the refractive index of a semiconductor cavity can be switched via the excitation of free - carriers in the semiconductor @xcite . however , the switching speed in such schemes is material dependent and limited by the recombination dynamics of the excited carriers . on the other hand ,
the refractive index of a semiconductor cavity can also be changed with the electronic kerr effect .
the electronic kerr effect is , in terms of speed , the ultimate way for ultrafast switching due to its material independent and instantaneous response nature @xcite . yet
, the excitation of relatively slow free carriers has to be avoided to accomplish a positive refractive index change with the electronic kerr effect , since free carriers lead to an opposite change of refractive index @xcite .
the main challenge is therefore to find a range of parameters where solely the kerr effect controls the optical properties of the cavity . in this work
we employ the instantaneous electronic kerr effect to switch the resonance frequency of a semiconductor planar microcavity in the original telecom band within 300 fs as a function of pump - pulse energy . using two light sources that provide pump and probe pulses we observe and analyse the competition of free carrier induced index changes and the electronic kerr effect in a switched cavity .
we have developed an analytical model which predicts this competition in agreement with our experimental data .
our model is developed for nondegenerate light sources and thereby can explain the effect of the cavity enhancement and the intensity of each source in the switching of a cavity resonance .
our experiments are performed on a planar microcavity grown by means of molecular - beam epitaxy .
the cavity resonance is designed to occur at @xmath1 in the original ( _ o _ ) telecom band .
[ yuce - f1 ] ( a ) shows a scanning electron micrograph of the sample .
the sample consists of a gaas @xmath2-layer ( @xmath3 ) sandwiched between two bragg stacks made of 7 and 19 pairs of @xmath4-thick layers of nominally pure gaas ( @xmath5 ) and alas ( @xmath6 ) , respectively and positioned on a gaas wafer .
the storage time of the probe photons in the cavity is deliberately reduced by decreasing the reflectivity of the top mirror of the cavity .
this leads to faster switching rates while at the same time reducing free carrier excitation due to a reduced field enhancement in the cavity .
-layer is indicated with white arrows and is sandwiched between two gaas - alas bragg stacks .
the gaas substrate is visible at the bottom .
the gaas layers appear dark grey , while the alas layers appear light grey .
( b ) measured ( black symbols ) and calculated ( red line ) reflectivity spectra of the microcavity .
the stopband of the bragg stacks extends from @xmath7 to @xmath8 .
fabry - prot fringes are visible on both sides of stop band . within the stopband a narrow trough at @xmath9 ( @xmath10 ) indicates the cavity resonance . from the linewidth ( @xmath11 , full width at half maximum ) of the cavity resonance we derive a quality factor @xmath12 corresponding to a cavity storage time of @xmath13 .
the calculations are performed with a transfer matrix model.__,width=8 ] ) .
two - photon absorption is largely suppressed by the judicious tuning of the pump and the probe frequencies relative to the semiconductor bandgap energy ( c ) schematic energy diagrams for all possible three - photon processes in the experiment that may result in free carrier generation ( see eq .
[ 5threephotondensity]).__,width=8 ] figure [ yuce - f1 ] ( b ) shows the measured and the calculated reflectivity spectrum of the microcavity .
the reflectivity spectrum of the cavity is measured with a setup consisting of a supercontinuum broad - band white - light source and a fourier - transform interferometer with a resolution of @xmath14 ( biorad fts6000 ) .
it can be seen that the stopband of the bragg stack extends from @xmath7 to @xmath8 ( @xmath15 ) . on both sides of the stopband fabry - prot fringes
are visible due to interference of the light reflected from the front and the back surfaces of the sample . inside the stopband
, a narrow trough indicates the cavity resonance at @xmath16 ( @xmath17 ) .
the resonance frequency of the switched cavity is determined by fitting a lorentz function to the resonance trough over a limited frequency range .
the relatively high reflectivity of the resonance minimum ( @xmath18 ) is a result of the asymmetric cavity design . from the linewidth ( @xmath11 , full width at half maximum ) of the cavity resonance we derive a quality factor @xmath12 corresponding to a cavity storage time of @xmath13 . a versatile setup described in ref .
@xcite is used to kerr - switch our microcavity .
the setup is shown in fig .
[ yuce - f2](a ) and consists of two independently tunable optical parametric amplifiers ( opa , light conversion topas pumped by a 1 khz oscillator ) that are the sources of the pump and probe beams .
the pulse duration of both opas is @xmath19 .
the time delay @xmath20 between the pump and the probe pulse is set by a delay stage with a resolution of @xmath21 .
the reflected signal from the cavity is detected with a nitrogen cooled ingaas line array detector spectrometer .
the measured transient reflectivity contains information on the cavity resonance during the cavity storage time and it should thus not be confused with the instantaneous reflectivity at the delay @xmath20 .
the measured transient reflectivity is a result of the probe light that impinges at delay @xmath20 which circulates in the cavity during the storage time and then detected which is integrated due to the relatively slow response time of the detector @xcite .
the cavity is switched with the electronic kerr effect by judicious tuning of the pump and the probe frequencies relative to the semiconductor bandgap @xcite .
the probe frequency ( @xmath22 ) is set by the cavity resonance in the telecom range while the pump frequency is centered at @xmath23 @xmath24 to suppress nondegenerate two - photon absorption ( @xmath25 ) see fig .
[ yuce - f2](b2 ) .
furthermore , the energy of the pump photons is chosen to lie below half of the semiconductor band gap energy ( @xmath26 ) , see fig .
[ yuce - f2](b1 ) , to avoid two pump - photon absorption .
the excitation of free carriers is also suppressed by choosing a low probe pulse energy ( @xmath27 ) while the average pump pulse energy is varied between @xmath28 .
the pulse energies are determined from the average laser power at the sample position and converted to peak power assuming a gaussian pulse shape .
the pulse energies are given per square micron since the switching of a cavity resonance has interesting prospects for miniature cavities with footprints in the micron range @xcite .
the pump beam has a larger gaussian focus ( @xmath29 ) than the probe beam ( @xmath30 ) to ensure that only the central flat part of the pump focus is probed and that the probed region is spatially homogeneously pumped .
the judicious selection of the pump - probe powers and frequencies enabled the instantaneous kerr switching at elevated frequencies including the telecom band .
figure [ yuce - f3 ] shows the transient reflectivity spectra for three different pump - probe time delays @xmath20 . at @xmath31
the probe pulse arrives earlier than the pump pulse .
hence , the measured spectra shows the unswitched transient reflectivity of the cavity with a resonance at @xmath32 . at temporal overlap of pump and probe pulses ( @xmath33 )
the cavity resonance frequency has red - shifted to @xmath34 indicating an increase of the refractive index . at positive delays ( @xmath35 ) , where the pump pulse arrives earlier than the probe pulse
, the cavity resonance is measured at @xmath36 .
thus the resonance frequency at positive delays has returned to the same frequency as the unswitched resonance frequency at negative delays .
the simultaneous observation of a red - shift of the cavity resonance only at pump - probe overlap and of the identical cavity resonances at positive and negative delays confirms that the cavity resonance is mainly switched by the electronic kerr effect , and not by the free carriers . pump pulse energy .
the orange curves show the fit to the cavity resonance from which the cavity resonance frequency ( @xmath37 ) is determined as the minimum.__,width=8 ] ) between pump and probe at different pump pulse energies .
the resonance frequency red - shifts due to the instantaneous electronic kerr effect only at temporal overlap ( @xmath38 ) of pump - probe ( shaded with bright color ) .
the blue shift of the cavity resonance due to free carriers is observed when the pump pulse energy is increased ( shaded with dark color ) .
the dotted horizontal line shows the unswitched resonance frequency at @xmath39 _ , width=8 ]
figure [ yuce - f4 ] shows the resonance frequency versus time delay at three different pump - pulse energies .
figure [ yuce - f4 ] is obtained from spectra similar to those shown in fig .
[ yuce - f3 ] . when the sample is pumped at @xmath40 the resonance frequency red shifts by @xmath41 at @xmath42 .
we observe the dynamic red - shift of the cavity resonance only at pump - probe coincidence within 300 fs , confirming the instantaneous switching of the cavity resonance frequency @xcite .
the red - shift of the resonance frequency induced by the electronic kerr effect increases to @xmath43 when the cavity is pumped at @xmath44 . at these power levels
the cavity resonance frequency at positive time delays returns to the same value as the unswitched cavity resonance . further increasing the pump pulse energy to @xmath45 results in an instantaneous shift of only @xmath46 although the sample is pumped with a three times higher pulse energy . at high pump energies
we also observe that the resonance frequency is blue shifted at positive time delays . at positive delays the refractive index decreases as a result of free carriers that remain excited for a much longer time ( about @xmath47 ) as has been observed before @xcite .
the carriers are excited by two- and three - photon processes as depicted in fig .
[ yuce - f2 ] ( b ) and ( c ) .
moreover , at high pump energies the cavity resonance frequency at @xmath48 is already blue shifted compared to the cold cavity resonance at low pump pulse energies , likely since the light is stored in the cavity up to @xmath31 which results in free carrier excitation by nondegenerate two- and three - photon absorption .
we conclude from fig .
[ yuce - f4 ] that the reversible and ultrafast cavity switching as a result of the electronic kerr effect occurs mainly at low pump pulse energies ( @xmath49 ) . in this regime ,
the switching speed is only limited by the cavity storage time and not by material relaxation properties .
the switching of the cavity can be achieved within 300 fs , which is only limited by the storage time of light in the cavity and not by extrinsic material properties . ) .
the orange curves show the fit to the cavity resonance from which the cavity resonance frequency ( @xmath37 ) is determined as the minimum , indicated as ticks.__,width=8 ] figure [ yuce - f5 ] shows the transient reflectivity spectra for three different pump - pulse energies at temporal ( @xmath33 ) and spatial overlap of the pump and the probe beams .
the cavity resonance frequency ( @xmath50 ) , shifts to a lower frequency ( @xmath51 ) if the sample is pumped at a low pump - pulse energy @xmath52 .
when the pump - pulse energy is increased to @xmath53 we observe that the instantaneously switched resonance frequency further red - shifts to @xmath54 .
however , when the pump - pulse energy is increased to @xmath55 the resonance frequency shifts to only @xmath56 showing less shift than the previous step . we conclude that at high pump - pulse energies a competition takes place between the electronic kerr effect that increases the refractive index and red - shifts the cavity resonance with the excited free carriers that decrease the refractive index and blue - shift the cavity resonance . error bar . at low pump - pulse energies ( bright shaded region )
we only observe electronic kerr effect since the resonance frequency decreases linearly with the pump pulse energy ( magenta dashed line ) .
the competing blue shift of free carriers is observed beyond @xmath57 and the region with both kerr and free carrier excitation is darker shaded .
the right ordinate shows the calculated refractive index change .
the red solid curve indicates the modelled index change as a function of the pump - pulse energy for general nondegenerate pump - probe light beams.__,width=8 ] figure [ yuce - f6 ] shows the shift of the instantaneously switched resonance frequency versus pump - pulse energy .
we observe that at low pump pulse energies the resonance frequency shift increases linearly with pump - pulse energy due to the positive refractive index change of the electronic kerr effect @xcite .
the linear increase of both the cavity resonance and the refractive index is physically reasonable since the nonlinear index change with the kerr effect is a product of the nonlinear susceptibility with the pump field squared ( see section modeling @xmath58 ref @xcite ) . beyond @xmath59
we observe a saturation and even a turn over of the resonance frequency shift versus pump - pulse energy . at high pump pulse energies , free carriers
are excited that reduce the refractive index , opposite to the kerr effect .
since the carriers with our settings of light frequencies can only be excited by two- and three - photon processes , the dependence of the refractive index ( and hence resonance frequency ) becomes nonlinear versus pump - pulse energy as is apparent from fig .
[ yuce - f6 ] .
we see in fig .
[ yuce - f6 ] that there is an apparent saturation in our experimental results .
our model does not show this saturation but a maximum .
since the pulse energies are limited in our experiments we can not observe a possible decrease of the refractive index at high pulse energies .
moreover , in our model we do not consider the plasma screening effect and the stark effect . from the linear slope at low pump energies
where free carriers are negligible , we derive a _ nondegenerate _
third - order nonlinear susceptibility @xmath60 for gaas at the strongly nondegenerate conditions @xmath22 and @xmath23 ( dashed line in fig . [ yuce - f6 ] ) .
the value that we find for @xmath0 agrees within an order of magnitude with degenerate values reported at @xmath61 @xcite .
the qualitative agreement between these different measurements is gratifying in view of the differences in the frequencies of the light sources @xcite .
in order to interpret the competition between the electronic kerr effect and free carrier effects at elevated pump - pulse energies , we have developed a model of the optical properties of the semiconductor cavity . here , we notably develop a model to include the nondegenerate three photon absorption . since only the two - photon absorption coefficient has been reported previously for the nondegenerate case @xcite , we chose to develop a model for nondegenerate case that can predict _ both _
two- and three - photon absorption cross sections . in our experiment , as a result of the cavity field enhancement the effect of the probe intensity becomes comparable to the pump intensity .
for this reason , we chose to develop a new model instead of using the existing degenerate three - photon absorption models @xcite to calculate the free carrier density as a function of both pump and probe intensities _
independently_. our model describes the refractive index change of a cavity resulting from both the electronic kerr effect and the free carriers in a nondegenerate pump - probe experiment .
in order to model the refractive index change induced by the electronic kerr effect we use the @xmath0 value determined from our experiments to limit the number of free parameters . for free carriers
the relevant @xmath0 has been theoretically described using the independent particle approximation @xcite .
the index change induced by free carriers that are excited by two ( fig .
[ yuce - f2](b ) ) and three ( fig .
[ yuce - f2](c ) ) photon absorption is calculated using the well - known drude model @xcite .
we can safely neglect the contribution of free carriers generated by one - photon absorption since both the pump and the probe photons are much less energetic than the band - gap energy of gaas .
we combine both the electronic kerr effect and the index change resulting from excited free carriers in eq .
[ 1index ] , which gives the refractive index change @xmath62 for a semiconductor optically switched in a general nondegenerate pump - probe experiment .
@xmath63}_{kerr } - \underbrace{{q^2 \over 2n_0\epsilon_0 m_{opt}^ * \omega_{pr}^2 } [ n_{eh}^{(2)}+n_{eh}^{(3)}]}_{free \ carriers } , \label{1index}\ ] ] here , @xmath0 is the third order nonlinear susceptibility , @xmath64 the vacuum permittivity , @xmath65 and @xmath66 the electric field of the probe in the cavity and of the pump pulses , respectively , @xmath67 the electron charge , @xmath68 and @xmath69 the frequencies of the probe and the pump beams , respectively , @xmath70 the optical effective mass of the free carriers , @xmath71 and @xmath72 the free carrier densities generated by two - photon and three - photon absorption , respectively .
[ 1index ] shows that the refractive index increases with the electronic kerr effect and decreases with the increasing density of free carriers . in eq .
[ 1index ] the refractive index change induced by the electronic kerr effect depends on the square of the electric fields of _ both _ the pump and the probe light ( the factor 2 for @xmath66 will become clear at eq . [ 3polarization ] ) . the electric fields in the instantaneous ( kerr ) part of eq .
[ 1index ] can be written in terms of cycle - averaged intensity using the relation @xmath73 .
in general , when describing the intensity - dependent refractive index the effect of the probe is neglected due to its smaller intensity compared to the pump @xcite . however , this is not necessarily the case for cavities due to the resonant field enhancement inside the cavity , which appears to be the case in our experiment . in general
the cavity enhancement is given by @xmath74 , where q is the quality factor of the cavity and r the reflectivity of the cavity mirrors @xcite .
the probe - pulse energy in our experiments is around @xmath75 before entering the cavity .
our cavity with q=390 enhances the probe field by @xmath76 times to @xmath77 so that it becomes non - negligible compared to the typical pump pulse energy ( @xmath78 ) in the cavity .
the resonant enhancement of the probe pulses by the cavity becomes even more important for high quality factor cavities @xcite , which might even bring the probe intensity beyond the level of the pump intensity . as a result , if the effect of probe light is neglected ( see eq .
[ 1index ] ) the usual pump intensity - dependent refractive index change will be incorrect especially for high quality factor cavities . in order to calculate the intensity - dependent refractive index for the general nondegenerate pump - probe case
we start by writing the total optical field as @xmath79 .
the general form of the total polarization of a material is described up to third order by @xmath80 due to the centrosymmetry of the gaas @xmath81 the total polarization of the material reduces to @xmath82 @xcite . taking the third power of the total optical field for nondegenerate pump - probe light and inserting it into the total polarization leads to @xmath83 which is the nonlinear polarization that influences the propagation of a beam of frequency @xmath68 .
the two - fold degeneracy factor in front of @xmath84 in eq .
[ 1index ] is due to the two - fold coefficient of the last term in eq .
[ 3polarization ] ( see appendix a for derivation ) . with this correction our model takes the cavity field enhancement into account and gives the appropriate solution for the intensity dependent refractive index for comparable intensities of nondegenerate pump - probe light .
having explained the kerr term of eq .
[ 1index ] we calculate the free carrier term of eq . [ 1index ] by calculating the free carrier densities as follows : @xmath85 , \label{4twophotondensity } \end{aligned}\ ] ] and @xmath86 .
\label{5threephotondensity } \end{aligned}\ ] ] the two- and three - photon free carrier densities ( @xmath71 and @xmath72 ) in eqs . [ 4twophotondensity ] and [ 5threephotondensity ] are calculated by multiplying the excitation rate @xmath87 with the number of atoms @xmath88 in the unit volume and the interaction time @xmath89 ( see appendix b and c for derivations of general case ) .
the relatively slow response time of the free carriers and the accumulation nature of the free carrier excitation will integrate in time and this will mask the ultrafast dynamics .
therefore , the intensities in eqs .
[ 4twophotondensity ] and [ 5threephotondensity ] are defined as @xmath90 , which differs from the instantaneous intensity term ( @xmath91 ) in the kerr term of eq .
[ 1index ] .
the limits of the time integral are given by the duration time of the excitation process , which is much longer than the excitation time of the carriers .
the pump interaction time is given by the pulse duration @xmath92 whereas the probe interaction time is given by @xmath93 due to the cavity which is in resonance with the probe light only . in eqs .
[ 4twophotondensity ] and [ 5threephotondensity ] , @xmath94 and @xmath95 are the density of final states and @xmath96 are the dipole transition moments associated with the resonance schemes depicted in fig .
[ yuce - f2](b ) and ( c ) .
the parameters used in our model are listed in table [ table1 ] .
.parameters used in our model . [ cols= " < , > , < , < , > " , ] [ table2 ]
as opposed to the derivation of three photon absorption coefficient ( @xmath97 ) that restricts itself to the simplified case of degenerate optical fields @xcite , we have here derived the two- and three - photon absorption cross sections for the general case of nondegenerate optical fields . since
only the two - photon absorption coefficient has been reported previously for the nondegenerate case @xcite we chose to derive a model for nondegenerate case that can predict _ both _
two- and three - photon absorption cross sections .
our approach holds the additional advantage of calculating the free carrier density as a function of both pump and probe intensities _
independently_. this feature allows us to extend refractive index changes for switched cavities with different quality factors q. we assumed cavities with resonance @xmath98 pumped at @xmath99 as in our experiment .
the pump pulse duration is taken as @xmath19 , whereas the probe pulse duration is set by @xmath93 since only the probe pulse is in resonance with the cavity and @xmath93 is inversely proportional to the quality factor q. figure [ yuce - f7 ] shows that the observed refractive index increase from the kerr effect can be revealed with low quality factor ( q=300 ) cavity up to @xmath100 pump pulse energy , similar to our experiment . for increasing quality factors
there is only a small increase in the refractive index due to the kerr effect before the free carriers decrease the index ( q=600 ) or even only a decreasing refractive index with increasing pump pulse energy ( q=1000 ) .
the less apparent kerr effect with the increasing quality factor is caused by the decreasing temporal overlap of pump and probe as the probe pulse becomes much longer than the pump pulse ( @xmath101 ) @xcite .
in fact , for high quality factor cavities during a longer fraction of the probe pulse there is no pump light , as a result no kerr switching occurs for this time duration . as a consequence , high quality factor cavities invite kerr switching with long pump pulses , but this defies the purpose of ultrafast optical switching .
interestingly , however , there is ( fig .
[ yuce - f7 ] ) also already kerr - induced refractive index increase for zero pump - pulse energy .
this effect is the result of the cavity enhanced probe light that already induces a kerr - shift . with increasing quality factor ,
the shift increases because of the increased probe - enhancement in the cavity .
however , due to the competing free carriers generated via degenerate two- and three- probe photon absorption the kerr induced positive shift does not scale linearly with the quality factor . we therefore conclude that cavities with shorter storage times can reduce the free carrier excitation which enables instantaneous switching of semiconductor cavities at the telecom range .
pumped at @xmath99 as in our experiment .
the pump pulse duration is taken as @xmath19 , whereas the probe pulse duration is set by @xmath93.__,width=8 ]
we demonstrate switching of a semiconductor microcavity within 300 fs at telecom wavelengths using the electronic kerr effect as a function of pump - pulse energy . we manage to measure the nondegenerate third order susceptibility ( @xmath0 ) of gaas using pump - probe experiment .
we show that the refractive index change induced by the electronic kerr effect can be increased to a certain extent that is limited by the increasing density of excited free carriers .
we show that the judicious tuning of the frequency of the driving fields relative to the band gap of the semiconductor decreases the number of free carriers and thereby increases the positive shift of the resonance frequency resulting from the electronic kerr effect .
our model quantitatively describes the frequency and the intensity dependence of nondegenerate switching with pump - probe experiment .
the realization and the understanding of the competition between the electronic kerr effect and the free carriers reveals the set of parameters using which the instantaneous electronic kerr effect can be utilized as the ultimate - fast way of all - optical switching . the refractive index change ( @xmath102 ) induced by the electronic kerr effect will result in a larger resonance frequency shift in comparison to the cavity linewidth with high - q cavities
however , we find that due to the larger field enhancements in high - q cavities the kerr effect will be hindered by the free carriers . on the other hand ,
if the incident probe pulse energy is further reduced then this competition in high - q cavities can be directed in favour of the electronic kerr effect . reducing the required pulse energy of the pump pulses will be achieved if the pump pulses are also resonantly enhanced by the cavity .
moreover , we note that the pump photons are not absorbed in the electronic kerr effect , hence the pump pulses do not heat up the sample and thus they can be recycled to switch the cavity resonance again .
in nonlinear optics the intensity dependent refractive index for pump - probe experiments is generally described under the assumption of weak probe field @xcite . in experiments involving cavities however ,
as described in this paper , pump and probe fields may be of the same order due to the resonant enhancement of the probe field . in order to reveal the consequences of this ,
we derive in this appendix the intensity - dependent refractive index involving nondegenerate pump and probe fields .
the nonlinear refractive index can be described for large interaction times @xcite @xmath103 where @xmath104 is the weak field refractive index , @xmath105 the second - order index of refraction and @xmath106 the time average of the electric field .
the arguments of @xmath105 express that the result at frequency @xmath68 depends on both @xmath68 and @xmath69 , as we will see below . in order to calculate the intensity dependent refractive index for nondegenerate pump - probe light
we start with an optical field of the form @xmath107 so that @xmath108 by inserting eq . [ 8averagefield ] into eq .
[ 6index ] we rewrite the nonlinear index in terms of the pump and probe fields @xmath109 the general form of the total polarization of a material is described up to third order by @xcite @xmath110 due to the centrosymmetry of gaas the total polarization reduces to @xmath82 @xcite .
taking the total optical field to the third power ( eq . [ 7field ] ) and inserting it into the total polarization leads to @xmath111 which is the nonlinear polarization that influences the propagation of a beam of frequency @xmath68 .
we introduce an effective nonlinear susceptibility in eq .
[ 11polarization ] given by @xmath112 we note that it is generally true that @xcite @xmath113 and by inserting eq . [ 9index ] and [ 12effectivetensor ] into eq .
[ 13index ] we get @xmath114 making the reasonable assumption that @xmath115 and by equating the terms of the same order on each side of eq .
[ 14indexsuscep ] we find the relation between the linear and nonlinear refractive indices and the relevant susceptibilities as follows : @xmath116 @xmath117 inserting eq .
[ 16index ] into eq .
[ 6index ] results in the nonlinear index @xmath118 equation [ 17indexfinal ] shows that the refractive index change induced by the electronic kerr effect depends on the square of the electric fields of _ both _ the pump and the probe light . for low values of @xmath119
, we get the usual expression for two degenerate beam case given in @xcite . in the text we use @xmath65 instead of @xmath119 since
only the probe pulse is in resonance with the cavity , which modifies the probe pulse duration and the intensity of the probe pulse .
however , in the appendix the equations are derived for a general case where the probe pulse is non - resonant .
following the derivation of refractive index change induced by the electronic kerr effect we derive the two - photon absorption rate @xmath120 using a perturbation approach to solve schrdinger s equation for nondegenerate applied optical fields .
we start with a two - level system to calculate the two - photon absorption rate and later we introduce density of states in order to mimic a semiconductor . in our derivation
we choose to explicitly write out all terms instead of using the permutation operator for the probability amplitude as in @xcite . in this way we can calculate the absorption rate as a function of both pump and probe intensities as an extension beyond the textbook @xcite .
we start by writing the time - dependent schrdinger equation in the presence of a time - dependent interaction potential @xmath121 .
we then use the standard perturbation analysis as described in ref .
@xcite to get : @xmath122 where @xmath123 is the probability amplitude of sate m with n interaction order and @xmath124 are the matrix elements of interaction hamiltonian @xmath125 .
we first calculate the linear absorption term , hence we set n=1 .
we assume that in the absence of any applied electric field the atoms are in the ground state _
( see fig .
[ yuce - f2 ] for the energy levels ) so that @xmath126 and @xmath127 for @xmath128 at all times _ t _ @xcite .
we then write @xmath129 as : @xmath130 , \label{19potential1}\ ] ] where @xmath131 is the transition dipole moment between states m and g. inserting eq .
[ 19potential1 ] into eq .
[ 18schrodinger ] gives @xmath132 .
\label{20schlevel1}\end{aligned}\ ] ] we drop the terms with @xmath133 since they describe the process of stimulated emission .
the neglect of the second terms is known as the rotating wave approximation . to get the probability amplitude for linear absorption we integrate eq .
[ 20schlevel1 ] @xmath134 \nonumber \\ & = & \frac{\mu_{mg}e_{pr}}{\hbar ( \omega_{mg}-\omega_{pr})}[e^{i ( \omega_{mg } -\omega_{pr})t}-1]+\frac{\mu_{mg}e_{pu}}{\hbar ( \omega_{mg}-\omega_{pu})}[e^{i ( \omega_{mg } -\omega_{pu})t}-1 ] .
\label{21problevel1}\end{aligned}\ ] ] in order to get the probability amplitude @xmath135 for two - photon absorption ( see fig . [
yuce - f2](b ) for the energy levels ) we describe @xmath136 as : @xmath137 , \label{22potential2}\ ] ] we use eqs . [ 21problevel1 ] and [ 22potential2 ] in eq.[18schrodinger ] to get @xmath138 in eq .
[ 23schlevel2 ] we again use the rotating wave approximation and we omit the complex terms which describe the stimulated emission as has been shown in eq . [ 20schlevel1 ] .
furthermore , in eq . [ 23schlevel2 ]
we assume that single level _ m _ dominates the sum so that the sum disappears .
thus we get : @xmath139 \nonumber \\ & + & \frac{e_{pr } e_{pu}}{(\omega_{mg}- \omega_{pu})}[e^{i ( \omega_{mg}+\omega_{nm } -\omega_{pr } - \omega_{pu})t}-e^{i ( \omega_{nm}-\omega_{pr})t } ] \nonumber \\ & + & \frac{e_{pr } e_{pu}}{(\omega_{mg}- \omega_{pr})}[e^{i ( \omega_{mg}+\omega_{nm } -\omega_{pr } - \omega_{pu})t}-e^{i ( \omega_{nm}-\omega_{pu})t } ] \nonumber \\ & + & \frac{e_{pu}^2}{(\omega_{mg}-\omega_{pu})}[e^{i ( \omega_{mg}+\omega_{nm } -2\omega_{pu})t}-e^{i ( \omega_{nm}-\omega_{pu})t}]\bigg ] .
\label{24schlevel2}\end{aligned}\ ] ] in eq .
[ 24schlevel2 ] we use the identity @xmath140 and we assume that the one - photon transition is highly non - resonant so that @xmath141 and @xmath142 . the terms with @xmath143 give the transient response of the process so that they can be dropped in the consideration of eq .
[ 24schlevel2 ] @xcite .
finally , we integrate eq .
[ 24schlevel2 ] up to time _
t _ and then multiply by @xmath144 to make the denominators look similar to the exponents : @xmath145}{\omega_{pr}(\omega_{ng}-2\omega_{pr})t } + \frac{e_{pu}^2 [ e^{i ( \omega_{ng}-2\omega_{pu})t}-1]}{\omega_{pu}(\omega_{ng}-2\omega_{pu})t } \nonumber \\ & + & \left ( \frac{e_{pr } e_{pu}}{\omega_{pu}}+\frac{e_{pr } e_{pu}}{\omega_{pr } } \right ) \frac { e^{i ( \omega_{ng}-\omega_{pr } -\omega_{pu})t}-1 } { ( \omega_{ng}-\omega_{pr } -\omega_{pu})t } \bigg ] .
\label{25problevel2}\end{aligned}\ ] ] we set @xmath146 so that eq .
[ 25problevel2 ] simplifies to : @xmath147 .
\label{26problevel2}\end{aligned}\ ] ] then the probability is @xmath148 .
\label{27problevel2}\end{aligned}\ ] ] in the following we analyse the terms inside the square brackets of eq .
[ 27problevel2 ] for long interaction times .
first we introduce functions f(t ) and g(t ) given as : @xmath149 @xmath150 \nonumber \\ & = & t^2[\frac{2 + 2cos(x - y)-2cosx-2cosy}{xy } ] .
\label{29functiong}\end{aligned}\ ] ] , which can be approximated as a dirac delta function .
( b ) the function g(t ) in eq .
[ 29functiong ] , which can be approximated as two - dimensional dirac delta function .
the absorption probability is proportional to the functions f(t ) and g(t).__,width=8 ] figure [ yuce - f8 ] ( a ) and ( b ) shows the the approximation of the functions @xmath151 as dirac delta functions for long interaction times , respectively .
the peak value of @xmath152 is @xmath153 when @xmath154 that is @xmath155 .
the width of the central peak is of the order of @xmath156 .
thus the area under the central peak is of the order of @xmath157 .
the function f(t ) can be expressed in terms of a dirac delta function for large t as @xcite : @xmath158 similarly we can also write @xmath159 in terms of dirac delta functions for large t as : @xmath160 we can see that ( fig .
[ yuce - f8 ] ) @xmath161 , if @xmath162 at the same time .
since @xmath163 then @xmath164 for nondegenerate light sources .
later , we consider that the presence of the delta function is somewhat unphysical .
instead we use the fact that the final state n is spread into a density of final states @xmath94 which is normalized such that @xcite @xmath165 and thus @xmath166 considering the fact that the first three terms inside square brackets in eq .
[ 27problevel2 ] have the same form as @xmath152 , we can apply the same procedure ( described from eq .
[ 28functionf ] to [ 33densityoflevels ] ) for these three terms .
moreover , due to the fact that @xmath164 for nondegenerate sources , the last six terms inside the square brackets in eq .
[ 27problevel2 ] become also zero since they have the form of @xmath159 . as a result , the probability to be in the upper level n simplifies to @xmath167 .
\label{34problevel2}\end{aligned}\ ] ] since the probability for an atom to be in the upper state seems to increase linearly with time , we can define a transition rate as @xcite @xmath168 the two - photon cross section can then be calculated via @xmath169 where @xmath170 is the intensity of the incident field in units of @xmath171 .
we can now calculate the density of free carriers from the transition rate using @xmath172 where @xmath88 is the number of interacting atoms per unit volume and @xmath89 the interaction time .
then we use the explicit forms of a , b , and c and we write the density of free carriers generated by two - photon absorption : @xmath173 .
\label{39twophotonappendix } \end{aligned}\ ] ] using eq .
[ 39twophotonappendix ] we can calculate the density of free carriers generated via two - photon absorption as function of pump and probe intensities and interaction times independently . in our calculations
we assume that the laser frequency is tuned to the peak of the two - photon resonance , so that @xmath174 @xcite where @xmath175 is the width of level _
n _ ( see table [ table1 ] )
[ 39twophotonappendix ] is calculated for the general case of two - photon absorption where all permutations of pump and probe exceed the bandgap energy .
one has to consider to leave out the non - resonant terms ( that do not excite free carriers ) to calculate the specific density of free carriers generated via two - photon absorption .
in order to calculate the probability amplitude @xmath176(t ) for three - photon absorption ( see fig . [
yuce - f2](c ) for the levels ) we follow the same steps as in appendix b but modified for the three - photon absorption process . here
, we describe @xmath177 as : @xmath178 .
\label{40potential3}\ ] ] we use eq . [ 25problevel2 ] and [ 40potential3 ] in eq .
[ 18schrodinger ] to get @xmath179 @xmath180 after we carry out the multiplications in eq .
[ 41schlevel3 ] we use @xmath181 for simplification and we drop the terms with @xmath182 since they give the transient response of the process and we also drop the terms describing the stimulated emission @xcite . then we integrate eq .
[ 41schlevel3 ] ( see appendix b for the similar steps ) .
we assume that both the one - photon and two - photon transitions are highly non - resonant so that @xmath183 . finally , we get @xmath184 . \label{42problevel3}\end{aligned}\ ] ] in eq .
[ 42problevel3 ] we set @xmath185 then eq . [ 42problevel3 ] simplifies to : @xmath186 .
\label{44problevel3}\end{aligned}\ ] ] then the probability is @xmath187
. \label{45problevel3}\end{aligned}\ ] ] if we analyze the terms inside the square brackets in eq .
[ 45problevel3 ] for large interaction times , following the same steps used in appendix b ( from eq .
[ 28functionf ] to eq .
[ 33densityoflevels ] ) , the probability to be in upper level _
l _ simplifies to @xmath188 .
\label{46problevel3}\end{aligned}\ ] ] we can define a transition rate as @xcite @xmath189 the three - photon cross section can then be calculated via @xmath190 where @xmath170 is the intensity of the incident field in the units of @xmath171 .
we can now calculate the density of free carriers from the transition rate using @xmath191 where @xmath88 is the number of interacting atoms in the unit volume and @xmath89 the interaction time .
then we write the density of free carriers generated by three - photon absorption by simply inserting the functions in eq .
[ 43xyz ] into eq .
[ 46problevel3 ] : @xmath192 .
\label{51threephotonappendix } \end{aligned}\ ] ] using eq .
[ 51threephotonappendix ] we are able to explicitly calculate the density of free carriers generated via three - photon absorption as function of pump and probe intensities and interaction times independently . in our calculations
we assume that the laser frequency is tuned to the peak of the two - photon resonance , so that @xmath193 @xcite where @xmath194 ( see table [ table1 ] ) is the width of level _
l_. eq . [ 51threephotonappendix ] is calculated for the general case of three - photon absorption where all the permutations of pump and probe exceeds the bandgap energy .
one has to consider to leave out the non - resonant terms to calculate the specific density of free carriers generated via three - photon absorption .
we thank allard mosk , and pepijn pinkse for stimulating discussions .
this research was supported by smartmix memphis , and the qswitch anr project ( to jmg ) .
this work is also part of the research program of fom , which is financially supported by nwo .
j. p. reithmaier , g. sk , a. lffler , c. hofmann , s. kuhn , s. reitzenstein , l. v. keldysh , v. d. kulakovskii , t. l. reinecke , and a. forchel , `` strong coupling in a single quantum dot semiconductor microcavity system , '' nature ( london ) * 432 * , 197200 ( 2004 ) .
d. englund , a. majumdar , m. bajcsy , a. faraon , p. petroff , and j. vukovi , `` ultrafast photon - photon interaction in a strongly coupled quantum dot - cavity system , '' phys .
lett . * 108 * , 09360414 ( 2012 ) .
g. ctistis , e. yce , a. hartsuiker , j. claudon , m. bazin , j. m. grard , and w. l. vos , `` ultimate fast optical switching of a planar microcavity in the telecom wavelength range , '' appl .
* 98 * , 16111413 ( 2011 ) .
p. j. harding , h. j. bakker , a. hartsuiker , j. claudon , a. p. mosk , j. m. grard , and w. l. vos , `` observation of a stronger - than - adiabatic change of light trapped in an ultrafast switched gaas - alas microcavity , '' j. opt .
b * 29 * , a1a5 ( 2012 ) .
j. l. jewell , s. l. mccall , a. scherer , h. h. houh , n. a. whitaker , a. c. gossard , and j. h. english , `` transverse modes , waveguide dispersion , and 30 ps recovery in submicron gaas / alas microresonators , '' appl .
lett . * 55 * , 2225 ( 1989 ) .
t. rivera , f. r. ladan , a. izral , r. azoulay , r. kuszelewicz , and j. l. oudar , `` reduced threshold all - optical bistability in etched quantum well microresonators , '' appl .
. lett . * 64 * , 869872 ( 1994 ) .
t. g. euser , p. j. harding , and w. l. vos , `` broadband sensitive pump - probe setup for ultrafast optical switching of photonic nanostructures and semiconductors , '' rev .
instrum . * 80 * , 07310419 ( 2009 ) .
a. hartsuiker , p. j. harding , y. nowicki - bringuier , j. m. grard , and w. l. vos , `` kerr and free carrier ultrafast all - optical switching of gaas / alas nanostructures near the three photon edge of gaas , '' j. appl
. phys . * 104 * , 08310517 ( 2008 ) .
g. ctistis , a. hartsuiker , e. van der pol , j. claudon , w. l. vos , and j .-
grard , `` optical characterization and selective addressing of the resonant modes of a micropillar cavity with a white light beam , '' phys .
b * 82 * , 19533017 ( 2010 ) . p. j. harding , t. g. euser , y. nowicki - bringuier , j. m. grard , and w. l. vos , `` dynamical ultrafast all - optical switching of planar gaas / alas photonic microcavities , '' appl
. phys . lett . *
91 * , 11110313 ( 2007 ) .
j. p. mondia , h. w. tan , s. linden , and h. m. van driel , `` ultrafast tuning of two - dimensional planar photonic - crystal waveguides via free - carrier injection and the optical kerr effect , '' j. opt .
b * 22 * , 24802486 ( 2005 ) . | we have performed ultrafast pump - probe experiments on a gaas - alas microcavity with a resonance near 1300 nm in the original " telecom band .
we concentrate on ultimate - fast optical switching of the cavity resonance that is measured as a function of pump - pulse energy .
we observe that at low pump - pulse energies the switching of the cavity resonance is governed by the instantaneous electronic kerr effect and is achieved within 300 fs . at high pump - pulse energies the index change induced by free carriers generated in the gaas start to compete with the electronic kerr effect and
reduce the resonance frequency shift .
we have developed an analytic model which predicts this competition in agreement with the experimental data . to this end
we derive the nondegenerate two- and three - photon absorption coefficients for gaas .
our model includes a new term in the intensity - dependent refractive index that considers the effect of the probe pulse intensity , which is resonantly enhanced by the cavity .
we calculate the effect of the resonantly enhanced probe light on the refractive index change induced by the electronic kerr effect for cavities with different quality factors . by exploiting the linear regime where only the electronic kerr effect is observed ,
we manage to retrieve the nondegenerate third order nonlinear susceptibility @xmath0 for gaas from the cavity resonance shift as a function of pump - pulse energy . |
the rashba spin - orbit coupling @xcite that exists in systems with axial symmetry , plays a key role in spintronics @xcite realized with two - dimensional ( 2d ) carriers in semiconductor heterostructures as it allows to manipulate the spin by a gate - controlled confinement potential .
spin - orbit ( so ) coupling mixes the spin states and removes the spin degeneracy for states with finite momentum . besides the rashba term caused by the asymmetry of the confinement , there exists also a so coupling due to the inversion asymmetry of the crystalline structure of the semiconductor bulk material ( dresselhaus term @xcite ) .
both types of so coupling combine to an anisotropic spin - splitting of 2d electrons , which has been analyzed by inelastic light scattering @xcite and play a role also in weak localization studies @xcite .
the intimate relation between spin splitting and spin relaxation , well - known for bulk material @xcite has found renewed interest for 2d electrons @xcite , furthered by the possibility to measure spin - relaxation times with monopolar optical orientation @xcite .
the zero - field spin splitting @xcite has to compete with the zeeman spin - splitting if a magnetic field is applied perpendicular to the plane of the 2d electron system .
this results in a fan chart showing characteristic crossings of landau levels , which in magnetotransport data are detected as beating of the shubnikov de - haas ( sdh ) oscillations @xcite . for hole sytems ,
the two so coupling mechanisms have been found to be responsible for anomalous sdh oscillations @xcite .
the structural asymmetry of the confinement can be tuned into a regime where the rashba so coupling dominates over the bulk inversion asymmetry @xcite . in spite of this current interest in the rashba so coupling and its relevance for spin - related transport in two - dimensional electron systems ( 2des )
it is surprising that there is so far no rigorous magnetotransport calculation which takes this coupling into account . here , we present fully quantum - mechanical calculations of the magnetoconductivity including the rashba so interaction .
the calculations are based on the evaluation of the kubo formula with an extension of the self - consistent born approximation ( scba ) by taking into account the electron spin degree of freedom .
our results show besides the known beats in the sdh oscillations an additional modulation connected with the crossing of landau levels .
this modulation , which is not seen in the density of states , arises as we take into account spin - conserving impurity scattering which is suppressed when the so coupled states are not degenerate .
it survives even at higher temperatures , when sdh oscillations have died out , and could serve , if experimentally detected , as another fingerprint of so interaction . in lateral superlattices @xcite , where a 2des is subjected to a periodic potential ,
there exist besides the sdh oscillations other @xmath0 periodic magnetotransport oscillations due to commensurability between cyclotron radius and lattice constant @xcite and due to the formation of a miniband structure @xcite . from our magnetotransport calculations for lateral superlattices with weak uni - directional ( 1d )
modulation we predict a splitting of these periods due to so coupling and calculate their dependence on its strength and carrier density . the paper is organized in the following way .
after this introduction we describe in the section [ second_section ] the energy spectrum and the eigenstates of a 2des with rashba so interaction . for the calculation of the conductivity
we simplify the energy spectum and introduce a constant spin - splitting model . in the section [ third_section ] the scba is extended by the spin degree of freedom to describe the scattering of spin - conserving impurities in a 2des with so interaction .
the effects of this extension are illustrated for a two level system , where it is shown that for non - degenerate so coupled states the scattering efficiency is suppressed . in section [ sec_conductivity ]
we present the conductivity and compare the cases with and without so coupling .
finally in the section [ fifth_section ] we show results for a system with so interaction and a 1d periodic modulation and study in the power spectrum of the magnetoconductivity the evolution of the characteristic periods with increasing so coupling .
the hamiltonian of a 2des ( in the @xmath1-plane ) realized in the lowest subband of a semiconductor heterostructure with effective mass @xmath2 , rashba so interaction due to the @xmath3-confinement with coupling constant @xmath4 , and zeeman term with effective @xmath5-factor @xmath6 , in a magnetic field @xmath7 , is given by @xmath8 where @xmath9 denotes the kinetic momentum and @xmath10 the pauli spin matrices , @xmath11 .
the spin ( _ up_/_down _ ) is quantized in @xmath3-direction .
the energy spectrum is isotropic and without the magnetic field , @xmath12 , it depends on the wavevector @xmath13 and is given by @xcite @xmath14 including the external magnetic field , the hamiltonian can be formulated with ladder operators @xmath15 with the parameters @xmath16 and @xmath17 , the cyclotron frequency @xmath18 and the magnetic length @xmath19 . in fig .
[ fig_one ] we show the spectrum of hamiltonian ( [ eqn_hamiltonian ] ) without magnetic field ( left ) and with magnetic field ( right ) for parameter values corresponding to inas ( @xmath20 , @xmath21 ) . in fig .
[ fig_one]a the situation without so coupling with spin - degenerate parabolic dispersion ( left ) and regular fan - chart of landau levels ( right ) is depicted . including spin - orbit coupling the picture of fig .
[ fig_one]b is obtained with the k - dependent splitting of the subband dispersion ( left ) and the characteristic crossing pattern of landau levels ( right ) . for this calculation
the so coupling parameter @xmath4 was choosen to be @xmath22evm close to the experimental values reported for inas samples @xcite .
the hamiltonian ( [ eqn_hamiltonian_ladder ] ) indicates that in the presence of so coupling the spin states , quantized in @xmath3-direction and used in the pauli representation , are no longer eigenstates . instead hamiltonian ( [ eqn_hamiltonian_ladder ] )
is diagonal for the states @xcite @xmath23 which can be classified by the helicity @xmath24 and the landau index @xmath25 .
@xmath26 are the eigenfunctions of the landau oscillator .
right_/_left _ states evolve with increasing so coupling from the spin _
up_/_down _ states , respectively , by the following choice of the coefficients : @xmath27 for the _ right _ states and @xmath28 for _ left _ states with @xmath29 and @xmath30 .
the energy eigenvalues of these states are @xmath31 in the following we use the notation @xmath32 where @xmath33 denotes the spin quantized in @xmath3-direction ( 1 : _ up _
, -1 : _ down _ ) and @xmath34 the helicity ( 1 : _ right _ , -1 : _ left _ ) .
the low temperature magnetoconductivity is determined by the electron states close to the fermi energy .
we will be interested mainly in the low - magnetic field regime where the conductivity is dominated by contributions from landau levels with @xmath35 and we may simplify the energy spectrum by replacing @xmath36 with @xmath37 . with the approximation @xmath38
we arrive at a model with constant spin - splitting @xmath39 and the coefficients of ( [ eqn_coeff_right ] ) and ( [ eqn_coeff_left ] ) take the forms @xmath40 and @xmath41 respectively , where @xmath42 . for this constant so coupling model
the energy eigenvalues are @xmath43 the energy spectrum of this model ( fig . [
fig_one]c ) consists of the two branches of _ right _ and _ left _ states which by our choice of @xmath44 are shifted by @xmath45mev .
the crossing of landau levels takes place for all levels at the same magnetic field .
thus our model preserves the two main effects of so coupling , the crossing of landau levels and the mixing of spin components _
for the calculation of the conductivity scattering has to be taken into account . the impurity - averaged green function of the system @xmath46 is given by the dyson equation @xmath47 the scattering is included by the selfenergy @xmath48 in self - consistent born approximation ( scba ) @xcite .
the matrixelements of the selfenergy @xmath49 , @xmath50 read in the basis of eigenstates ( [ eqn_rechtslinks_basis ] ) @xmath51 the kernel @xmath52 is given by the expression @xmath53 where @xmath54 denotes the fourier transform of the impurity potential @xcite .
we consider spin - conserving short range impuritiy scattering . by this choice we neglect magnetic impurities and so coupling with the scattering center , thus without rashba so coupling scattering is possible only between states with the same spin - quantum number . in the new basis ( [ eqn_rechtslinks_basis ] ) , including the rashba term , scattering between states with different helicity becomes possible . for @xmath55-scatterers
( [ eqn_gamma ] ) simplifies to @xmath56 where @xmath57 is connected through the relaxation time @xmath58 with the mobility @xmath59 in the case without magnetic field .
the kronecker symbols have the form @xmath60 and @xmath61 with @xmath62 and @xmath63 .
we restrict ourselves to the diagonal approximation in the spacial quantum numbers by considering only @xmath64 . in this approximation
the selfenergy reads @xmath65 with @xmath66 independent of @xmath67 , and the green function @xmath68 .
thus the scba selfenergy becomes @xmath69 with @xmath70 and @xmath71 where @xmath72 is the green function of the system without impurities . using the abbreviations @xmath73 , @xmath74 and @xmath75 the selfenergy
can now be calculated from @xmath76 the off - diagonal elements of the selfenergy will be neglected . the density of states for the _ left _ and _ right _ components are obtained by the trace over the spectral funktions @xmath77 .
@xmath78 here @xmath79 is the inverse of the matrix formed from the @xmath80 of eq .
( [ eqn_sigma_tautau ] ) for which @xmath81 .
for vanishing so coupling one has @xmath82 and the standard scba result is reproduced . for strong so coupling if @xmath83 , i.e. the landau splitting is smaller than the splitting induced by the so interaction , one has @xmath84 . due to the so interaction the landau levels of
different helicity cross as seen in fig .
[ fig_one]b and c. to demonstrate the influence of scattering around these crossing points we apply the described extension of the scba to a two level system with different spacings . the two states at the energies @xmath85 , @xmath86 have different helicity and evolve with increasing so coupling from the spin _
up_/_down _ states with energy @xmath87 and @xmath88 . in fig .
[ fig_two]a the situation without so interaction @xmath89 is shown . for decreasing energy difference @xmath90 the form of the selfenergy , calculated from eq .
( [ eqn_sigma_tautau ] ) , and the density of states , from eq .
( [ eqn_dos_tau ] ) , keeps unchanged .
selfenergies @xmath91 and density of states for a two - level system without ( a ) and with ( b ) so coupling for different separation of the up-/down- or right-/left - eigenstates in ( a ) and ( b ) , respectively . in ( b ) a mixing coefficient @xmath92 at the energies
@xmath85 and @xmath86 has been chosen . ] in contrast , in the case of so coupled states ( fig .
[ fig_two]b , with @xmath93 ) the impurity scattering adds to the selfenergy a contribution at the neighbouring state with opposite helicity .
if the states are not degenerate , both width and height of the imaginary part of the selfenergy are reduced as compared with fig .
[ fig_two]a , while for the degenerate levels the selfenergy and the density of states are identical with the situation of pure spin states . for
strongly so coupled states @xmath94 the height and width of the imaginary part of the selfenergy are reduced by @xmath95 , as can be seen from eq .
( [ eqn_sigma_tautau ] ) . as these parameters are inversely proportional to the relaxation time we conclude , that the scattering time is increased by the so coupling , away from the crossing points .
to summarize , the scattering strength depends on the level separation , i.e. on the distance from crossing points in fig .
[ fig_one ] . as
this distance varies with the magnetic field we expect a modulation of the scattering strength with the magnetic field .
based on the exact eigenstates of the constant so coupling model we evaluate the kubo formula @xmath96 to calculate the conductivity . here
the spectral function @xmath97 includes the impurity scattering in scba as described in sec .
[ third_section ] and the fermi distribution function yields the temperature average . the velocity is given by the equation @xmath98 $ ] , which results in @xmath99 besides @xmath100 we calculate also the thermodynamic density of states at the fermi energy @xmath101 . in fig .
[ fig_three ] both quantities are shown for the two temperatures @xmath102k and @xmath103k with the fermi energy determined from the constant electron density @xmath104 .
for comparison the classical high temperature limit of the scba for decoupled landau levels is given by the dashed line .
thermodynamic density of states at the fermi energy @xmath105 and longitudinal conductivity @xmath106 for the parameters of fig .
[ fig_one ] , electron density @xmath107m@xmath108 and mobility @xmath109m@xmath110vs at two temperatures @xmath111k ( black solid ) and @xmath112k ( grey solid ) .
the grey circless mark the expected position of maxima of the sdh oscillations without so coupling ; landau level crossings due to the so coupling are indicated by crosses .
the dashed line in the plot of the longitudinal conductivity is the classical limit of the scba when the landau levels are decoupled . ]
the effect of spin - orbit coupling is seen in @xmath113 as the beating pattern , well - known from measured sdh oscillations @xcite , while in @xmath106 it causes an additional modulation , which survives even at a higher temperature when the sdh oscillations are damped out .
the period of this modulation is determined by the crossing of landau levels ( marked in fig .
[ fig_three ] ) induced by the spin - orbit coupling .
the self - energy enters differently into @xmath114 and @xmath106 . in @xmath114
it leads to a modulated broadening of the landau levels which is washed out in the high - temperature limit , while in @xmath106 it acts in addition as a scattering time whose dependence on the level separation remains even at higher temperatures .
this is seen by comparing with the classical scba limit for decoupled landau levels ( dashed lines ) : the kubo formula ( [ eqn_cond_mumu ] ) yields a magnetoconductivity which is reduced away from the crossing points due to suppression of impurity scattering .
( note , that we are in the limit @xmath115 , where the drude conductivity is proportional to @xmath116 . ) to demonstrate the effect of scattering between states of different helicity in our extension of the scba , we show in fig .
[ fig_four ] the dependence of the conductivity at @xmath103k on the mixing coefficient @xmath117 by keeping the energy spectrum with spin - orbit splitting unchanged but varying @xmath118 , which otherwise is given by the coefficients @xmath119 . neglecting the scattering between states of different helicity @xmath120 , the crossing of landau levels
has no effect on the conductivity . for dominant spin - orbit coupling , @xmath94
, we see that the conductivity is reduced between crossing points of landau levels .
we can distinguish two situations ( top of fig .
[ fig_four ] ) : the states with different helicity are degenerate ( a ) and the scattering efficiency keeps unchanged ; when _
right_/_left_-states are not degenerate ( b ) the conductivity depends on @xmath117 .
( a ) sketch of landau level spectrum with ( a ) and without ( b ) degeneracy of states with different helicity .
( b ) longitudinal magnetoconductivity in dependence of the strengh of the so coupling expressed by the parameter @xmath121 for high temperature @xmath122k .
the parameters are those of fig .
[ fig_three ] and the dashed line is again the classical limit of the scba when the landau levels are decoupled . ]
longitudinal magnetoconductivity at fixed @xmath123 but with different @xmath124 at @xmath122k .
the other parameters are those of fig .
[ fig_three ] .
the dashed lines mark the classical limit of the scba with decoupled landau levels . ] in fig .
[ fig_five ] , we have varied the strength of the impurity scattering by changing the parameter @xmath125 . at the classical limit of the scba with decoupled landau levels ( dashed lines ) ,
we see that the conductivity rises with increasing @xmath125 .
the modulation of the conductivity due to the crossing of landau levels decreases , because the limit of decoupled landau levels can not be reached for large enough @xmath125 .
we have shown by a quantum mechanical calculation , that the so coupling causes the expected beats of the sdh oscillations .
in addition our results exhibit a new modulation of the magnetoconductivity which can be ascribed to a modification of impurity scattering in the presence of so coupling .
this could help in experiments to distinguish the effect of the so coupling from that of inhomogeneous electron densities which was invoked in ref .
@xcite to explain beatings in the sdh oscillations .
lateral semiconductor superlattices have proven to be well suited for studying the physics of bloch electrons in artificial periodic systems @xcite . a variety of oscillations have been observed in magnetotransport experiments for systems , in which the lattice constant @xmath126 is comparable with achievable magnetic lengths @xmath127 and both being much smaller ( at low temperature ) than the carrier mean - free - path . after having studied the sdh oscillations in the previous section we focus here on the influence of the rashba so interaction on the various magnetoconductivity oscillations due to the periodic modulation mentioned in the introduction . for the lateral superlattice with rashba so interaction
the hamiltonian is given by @xmath128 we consider here a periodic modulation in @xmath129 direction described by @xmath130 it removes the degeneracy of the landau levels and is taken into account in evaluating the kubo formula together with the spin - conserving impurity scattering in the extension of the scba as for the homogeneous 2des . in fig .
[ fig_six ] the longitudinal magnetoconductivities @xmath106 and @xmath131 , in the direction of the modulation and perpendicular to it , respectively , are depicted for potential parameters @xmath132mev , @xmath133 nm , and two temperatures .
the fermi energy was fixed to @xmath134mev , much larger than the amplitude of the periodic potential , which defines the weak modulation case .
the mixing coefficient was fixed to the maximum value @xmath135 .
different types of oscillations can be identified of which at the higher temperature ( @xmath136k ) only the commensurability oscillations and the modulation due to so coupling survive
. the periods ( seen in the @xmath137k traces ) can be quantified in simplified models .
following onsager @xcite the sdh periods are @xmath138 where the appearance of two fermi contours due to the spin - splitting is accounted for @xcite .
the period of the commensurability oscillations in given by @xcite @xmath139 where again the spin - split fermi contours are considered .
the periods reflecting the formation of the miniband structure due to the periodic potential @xcite are quantified by @xmath140 where @xmath141 is the cross section of the modified fermi contour given in ref .
again the spin - splitting due to so coupling is to be considered and eq .
( [ eqn_onsager ] ) yield two periods for the fermi cross sections @xmath142 finally we may conclude from the eigenvalues of eq .
( [ eqn_eigenvaluesrl ] ) for @xmath35 a period of @xmath143 for the modulation connected with the crossing points of the landau levels at the fermi energy . in order to analyze all these oscillations
we take the power spectrum of the differences @xmath144 , @xmath145 , of the calculated longitudinal conductivities at @xmath146 and @xmath147k .
it is shown in dependence on the strength of the so coupling ( fig . [ fig_seven ] ) and on the fermi energy @xmath148 ( fig .
[ fig_eight ] ) together with the periods predicted from eqs .
( [ eqn_d1db_sdhpm])-([eqn_sod1db ] ) . in fig .
[ fig_seven ] we see that longitudinal magnetoconductivities @xmath106 and @xmath131 for a lateral superlattice with uni - directional modulation in @xmath129-direction .
( modulation amplitude @xmath132mev , lattice constant @xmath133 nm , fermi energy @xmath149mev , mobility @xmath150m@xmath110vs and inas effective mass @xmath151 . the so coupling is @xmath152evm and the mixing coefficient @xmath153 . ]
the absolute value of the fourier spectrum of the difference of the conductivities at @xmath137k and @xmath147k analogous to fig .
[ fig_six ] is shown in dependence on @xmath4 .
the arrows mark the contribution due to the crossing of landau levels .
the calculated fourier spectrum is compared with the models of eqs .
( [ eqn_d1db_sdhpm]),([eqn_d1db_copm ] ) , ( [ eqn_onsager ] ) and ( [ eqn_sod1db ] ) . ] with increasing so coupling the frequencies of all ( but one ) resolved 1/b periodic oscillations split and the maxima of the fourier transform follow the predictions of the simplified models .
this is not the case for the peaks showing up at the lowest @xmath154 values . these peaks , being due to the modulation resulting from our extension of the scba ( sec .
[ third_section ] ) , do not split and follow the analytic expressions of eq .
( [ eqn_sod1db ] ) . in fig .
[ fig_eight ] the oscillation period is shown in dependence on the fermi energy . the absolute value of the fourier spectrum of the difference of the conductivities at @xmath137k and @xmath147k analogous to fig .
[ fig_six ] is shown in dependence on the fermi energy @xmath148 .
the arrows mark the contribution due to the crossing of landau levels .
comparison with eqs .
( [ eqn_d1db_sdhpm]),([eqn_d1db_copm ] ) , ( [ eqn_onsager ] ) and ( [ eqn_sod1db ] ) is shown by straight lines . ]
again the maxima of our full quantum - mechanical calculation follow the predictions of the simplified models of eqs .
( [ eqn_d1db_sdhpm])-([eqn_sod1db ] ) .
we have calculated the magnetoconductivity for a homogenous 2des with spin - orbit interaction including a non - trivial extension of the scba which takes into account the spin - degree of freedom .
the crossing of landau levels with different helicity , which is induced by the spin - orbit interaction , manifests itself in an additional modulation of the conductivity .
it is due to a modification of impurity scattering in the presence of so coupling .
this signature could be used - besides the beating of sdh oscillations - for the experimental verification of so coupling .
further we have investigated the influence of so interaction on the magnetoconductivity oscillations in 1d lateral superlattices .
the prediction of the frequencies by simple models and the numerical calculations are in agreement and should motivate an experimental verification . | we present magnetotransport calculations for homogeneous two - dimensional electron systems including the rashba spin - orbit interaction , which mixes the spin - eigenstates and leads to a modified fan - chart with crossing landau levels .
the quantum mechanical kubo formula is evaluated by taking into account spin - conserving scatterers in an extension of the self - consistent born approximation that considers the spin degree of freedom .
the calculated conductivity exhibits besides the well - known beating in the shubnikov - de haas ( sdh ) oscillations a modulation which is due to a suppression of scattering away from the crossing points of landau levels and does not show up in the density of states .
this modulation , surviving even at elevated temperatures when the sdh oscillations are damped out , could serve to identify spin - orbit coupling in magnetotransport experiments .
our magnetotransport calculations are extended also to lateral superlattices and predictions are made with respect to @xmath0 periodic oscillations in dependence on carrier density and strength of the spin - orbit coupling . |
our goal is to identify functional properties of nodes based on the network structure .
connection between network structure and its functionality is important , many attempts were made to find functional signatures in the network structure , such as @xcite , for a review see @xcite . as tagging network nodes and edges with functional attributes
depends on external information and is not a completely unique procedure , the original problem needs reformulation which is tractable with graph - theoretical tools .
the function real - world networks perform constrains their structure . yet , one often has more detailed information about the network structure than about the functions it may perform .
we focus on systems , either natural or artificial , which process signals and are comprised of many interconnected elements . from a signal processing point of view
, global information about network structure is encoded in the shortest paths , i.e. if signal processing is assumed to be fast , most of network communication is propagated along the shortest paths . therefore global and local properties of shortest paths are relevant for understanding organisation of the signal processing in the system represented with a suitable network . during signal transmission ,
signals are being spread and condensed in the nodes , as well as along network edges .
we have previously shown @xcite that in case of cerebral cortex , using a simplified version of the convergence degree ( cd ) , it was possible to connect structural and functional features of the network .
in complex networks , signal processing characteristics are also determined by the level of network circularity ( which in biology and especially neural science is known as reverberation , for obvious reasons ) .
possibility to go around _ chordless _ circles necessitates simultaneous quantification of signal condensing , spreading along network edges and edge circularity .
here we generalise edge convergence and divergence @xcite , and take into account the existence of circles in the network , treating their effects separately from the effect of branching .
for that reason we refine the definition of edge convergence and introduce the overlapping set of an edge , both notions are to be defined in a precise manner later in the text .
our approach may be viewed as generalisation of in- , out and strongly connected components of a graph to the level of network edges .
notions introduced have an extra gain , they help clarifying the otherwise murky notion of network causality
. the functional role of a node in a network is defined by the amount of information it injects to or absorbs from the system , or passes on to other nodes . in case of real - world networks
we test our findings using external validation , given the existing body of knowledge about each specific network .
we illustrate the advantage of edge - based approach with the case of strongly connected graphs , where edge - based measures offer deeper understanding of signal processing and transmitting roles of nodes than an analysis which concentrates solely on nodes and their properties .
measures we work with are applicable to networks of all sizes , there is no assumption about `` sufficient '' network size .
more precisely , networks we work with can be small , and applicability to large networks is limited only by the computational capacity needed to find all shortest paths in the network .
the semantics of our approach is tailored to explain signal flow , though our methodology is applicable to directed networks in general . in cases of information processing , regulatory , transportation or any other network
the appropriate semantics of the approach has to be given . in section [
sec : notion ] we introduce the notions of convergence degree and overlapping set , in section [ sec : nrr ] we define the flow representation , in section [ sec : res ] we analyse four real - world networks and discuss signal transmission , processing and control properties of the small - world networks .
we compute cd - s and ( nontrivial ) overlap probability distributions for three model networks . in the last section
we discuss our results and draw conclusions .
convergence degree was introduced in @xcite for the analysis of cortical networks and was applied to some random networks @xcite .
we modify the measure introduced therein , in order to capture the structure of shortest paths in a more detailed way .
we will discuss both global and local properties of the shortest paths , relevant notions will be distinguished with self explanatory indices @xmath0 and @xmath1 respectively .
let @xmath2 be the set of all the shortest paths in the graph @xmath0 .
for any edge @xmath3 we can choose a subset @xmath4 comprised of all the shortest paths which contain the chosen edge @xmath5 .
@xmath4 uniquely determine two further sets : @xmath6 the set of all the nodes from which the shortest paths in @xmath4 originate , and @xmath7 the set of all the nodes in which the shortest paths in @xmath4 terminate . by definition
we assume that node @xmath8 is in @xmath6 and node @xmath9 is in @xmath7 .
we define a third set , @xmath10 , the intersection of @xmath11- and @xmath12 sets and call it the overlapping set .
we note that @xmath6 ( @xmath7 , respectively @xmath13 ) is the edge - level equivalent of the in - component ( out - component , respectively strongly connected component ) of the directed network , introduced in @xcite and later refined by @xcite .
notions relevant for understanding the convergence degree and overlapping set are shown in figure [ fig : cd1 ] . .
global sets are displayed as shaded regions , local sets are comprised of first in - neighbours of node @xmath14 and first out - neighbours of node @xmath15 inside the shaded regions , with the exception of node @xmath0 , which is contained in the local and global overlap of @xmath16 and @xmath17 .
note the omition of points @xmath18 and @xmath19 from the global input and output sets . ] from the perspective of the chosen edge , the whole network splits to two , possibly overlapping sets , both of which have rich structure .
shortest paths induce natural stratification on the set @xmath6 , nodes at distance 1 , 2 and so on from the node @xmath8 are uniquely determined .
points at distance @xmath20 from the tail form the @xmath20-th stratum of @xmath6 .
each point in the @xmath20-th stratum is a tail of an edge with a head in the @xmath21-th stratum .
edges connecting @xmath20-th stratum with any stratum @xmath22 are prohibited .
edges from the @xmath11 strata to the @xmath12 strata are prohibited , since those would alter the shortest paths between the sets .
the set @xmath7 is stratified in a similar fashion .
points in the intersection of @xmath6 with @xmath7 inherit both stratifications .
stratification of @xmath23 and @xmath24 sets is illustrated in figure [ fig : strata ] .
, output strata are labelled with indices @xmath25 and overlap strata have double indices @xmath26 .
examples of prohibited edges are shown with dashed lines , necessary edges are shown with full line .
strata @xmath27 and @xmath28 are connected with the edge itself and they do not overlap . ] local versions of these sets are defined as follows : @xmath29 is the set of all the first predecessors of the node @xmath8 , while @xmath30 is the set of first successors of the node @xmath9 . when indices @xmath0 or @xmath1 are omitted , either is used .
if the graph has circles , @xmath11 and @xmath12 sets may overlap , thus it makes sense to introduce strict @xmath31 and @xmath32 sets , which are defined as follows : @xmath33 @xmath11 , @xmath12 , @xmath31 and @xmath32 are generalisations of the notion of first predecessors and successors of a node , and accordingly , cardinalities of these sets are generalisations of the in- and out - degrees of nodes .
we note that global and local versions of the @xmath11 , @xmath12 and overlapping sets are two extremes of two set families defined as follows .
let @xmath34 be the set of points from which paths at distance less or equal to @xmath35 from the point @xmath8 begin , analogously let @xmath36 be the set of points at which paths at distance less or equal to @xmath37 from the point @xmath9 terminate .
the two sets are balls centred at @xmath8 and @xmath9 with radii @xmath35 and @xmath37 . instead of balls , one may consider the surfaces of the balls , in which case points at distances @xmath35 and @xmath37 are considered .
the global @xmath11-set is thus @xmath38 , whilst the local @xmath11-set corresponds to points at surfaces with radii 1 , @xmath39 .
the notion of strict in- , out- and overlapping sets is important for understanding causality relations in network systems .
global signal flow through an edge @xmath5 induces separation of network nodes into four classes : 1 .
@xmath40 , in which are the causes of the flow .
@xmath41 , in which the effects of flow are manifested .
3 . the overlap , whose elements represent neither cause nor effect . relation between elements in the overlap
is often described as circular- or network causality .
points which are not members of @xmath42 form the remaining , fourth category which has no causal relationship with the signal flowing through the given edge .
we stress that for a generic graph no such partition is possible based on node properties .
e.g. if we tried to define analogous notions based on node properties , all analogue node classes would coincide for the case of strongly connected graphs .
the @xmath11 and @xmath12 sets would coincide , and all distinction between different node classes would have been lost . for each edge
we define three additional measures , namely the relative size of the strict in - set @xmath43 , the relative size of the strict out - set @xmath44 , and the relative size of the overlap between in - set and out - set @xmath45 , as follows : @xmath46 where @xmath47 denotes the cardinality of the set @xmath48 .
note that equation [ eq : uc3 ] is the jaccard coefficient @xcite of the @xmath49 and @xmath50 sets .
it is possible to generate networks which have edges with large global overlaps , one simply adds randomly a small number of edges to an initial oriented circle .
this example helps understanding the meaning of ( possibly large ) global overlaps : they are characteristic of edges in chordless circles . more precisely , for and edge to have a nonempty overlapping set it is necessary , but not sufficient , to be on a chordless circle of length at least three .
we illustrate this by an example . in the graph shown in figure
[ fig : circ ] , the only edge with nonempty overlapping set is @xmath51 , with @xmath52 .
@xmath51 is on the chordless circle ( 3,1,2,3 ) , whilst the edges @xmath53 and @xmath54 on the same chordless circle have zero overlapping sets .
local overlaps are related to the clustering coefficient of the graph , since they define the probability that the vertices in the neighbourhood of a given vertex are connected to each other .
overlap represents global mutual relationship and a measure of dependence ( in terms of chordless circles ) between @xmath11- and @xmath12 sets .
this dependence is inherent in the network structure .
large jaccard coefficient of the @xmath49 and @xmath50 sets is not detectable with edge betweenness , as it may obtain large values for edges with non - overlapping sets .
the edge convergence degree @xmath55 of the edge @xmath5 is defined as follows : @xmath56 note that the definition of cd uses the normalised sizes of the strict @xmath11- and @xmath12-sets to make the measure independent of the network size .
furthermore , this formula is related to the complement of the jaccard coefficient ( denoted as @xmath57 ) of the @xmath11- and @xmath12-sets , or equivalently to their normalised set - theoretic difference , thus connecting the cd to information theoretical quantities .
the following inequality is obvious : latexmath:[\[\begin{aligned } directionality of the edge gives meaning to cardinality substraction , as @xmath11 and @xmath12 sets can be distinguished .
if the cd value is close to one , the signal flow through the edge is originating from many sources and terminating in very few sinks , while cd values close to -1 indicate flow formed of few sources and many sinks .
this property justifies rough division of edges according to their cd properties to convergent ( condensing ) , balanced and divergent ( spreading ) .
an oriented circle with at least three nodes has the maximum possible global overlap for each edge , while the absolute value of the global @xmath59 is the smallest possible , in accordance with the inequality ( [ eq : cd_le_ovl ] ) .
we note that cd in an oriented chain monotonously decreases along the chain , whilst the overlap is zero along the chain .
this simple example again illustrates how cd and overlap are sensitive to the network topology .
applicability of the convergence degree is limited by the following facts .
definition of convergence degree makes sense only if not all connections are reciprocal , stated otherwise if there is a definite directionality in the network
. if every connection is reciprocal , the network may be considered unoriented .
for fully reciprocal networks , the @xmath11 and @xmath12 sets would coincide .
second , convergence degree makes sense for a network which is at least weakly connected .
since the number of edges exceeds the number of nodes in a typical connected network , and in many cases we are interested in the role of individual nodes , it is desirable to condense the our primarily edge - based measures to a node - centric view .
the condensed view should reveal several features of interest : local vs global signal processing properties of network nodes , directionality of the information , i.e. whether we are interested in the properties of the incoming or outgoing edges , the third aspect is the statistics , i.e. total or average property of the edges , and finally we may choose edges according to the sign of their cd . condensing
the information about overlapping sets follows the same lines , with the exception of the sign .
we proceed by an example and introduce the following six quantities defined for each node @xmath8 .
let @xmath60 denote the sum of all incoming negative local convergence degrees divided by the node s in - degree , and let @xmath61 denote the sum of all incoming positive convergence degrees divided by the node s in - degree , i.e. @xmath60 is the average negative inwards pointing local cd of the node @xmath8 . in a similar way we can also define @xmath62 and @xmath63 for outgoing convergence degrees . for clarity
we give formulae for @xmath60 and @xmath62 . @xmath64 and @xmath65 denote in - degree and out - degree of the node @xmath8 , @xmath66 is the unit step function continuous from the left .
@xmath67 denotes the first in - neighbours of the node @xmath8 , the analogous notation @xmath68 is selfexplanatory .
@xmath69 we also define @xmath70 , the sum of all incoming local overlaps and @xmath71 , the sum of all outgoing local overlaps each being normalised with the corresponding node degree .
@xmath72 factors before the sums serve normalisation purposes , each @xmath73 should have a value within the @xmath74 $ ] interval .
these quantities are average local cd - s and relative overlaps corresponding to each node .
one is also interested in the total of the in- and out pointing edges of a given cd sign , and define the corresponding version of the node - reduced convergence degree . for normalisation purposes the sums in @xmath75 s are divided by @xmath76 , the maximal possible number of the outgoing ( incoming ) connections a node can have , where @xmath77 denotes the number nodes in the network .
thus , using the quantities @xmath78 and @xmath79 one can construct four different cd flow representations of a network , namely @xmath80 , @xmath81 , @xmath82 and @xmath83 .
the incoming node - reduced cd values are understood as coordinates of the @xmath84 axis , while the outgoing cd values are interpreted as the coordinates of the @xmath85 axis . in order to display overlaps together with the convergence degrees in a single figure ,
overlaps are treated as the coordinates of the @xmath86 axis , the incoming overlaps being positive and the outgoing understood negative .
each point is represented in each octant of the flow representation .
the points in the @xmath87 plane are not independent , given the values in the diagonal quadrants , the other two quadrants can be reconstructed with reflections . representation of graph nodes in the @xmath87 plane is related to the cd flow through the nodes in the following way .
the cd flow @xmath88 through the node @xmath8 is defined as follows : @xmath89 the first sum is equal to @xmath90 , where @xmath91 is the appropriate weight , whilst the second sum equals @xmath92 .
the flow can be rewritten as @xmath93 if the first difference on the right hand side of equation ( [ eq : flow_2 ] ) is large ( small ) , i.e. the representative point is close to the diagonal @xmath94 and is far from the origin in the top left ( bottom right ) quadrant , and the second difference is small ( large ) , i.e. the representative point is close to the diagonal @xmath94 and is far from the origin in the bottom right ( top left ) quadrant , the node @xmath8 is _ source _ ( _ sink _ ) of the cd flow .
analogously , the cd flow can be written as : @xmath95 where the two differences determine the router characteristics of the node @xmath8 . in this sense flow representation
is a means to independently study different components of the cd flow .
different circles may have common nodes , thus the overlap flow defines whether different circles passing through the given node have more common parts after of before the given node , i.e. whether a node is a source or sink of circularity .
precise meaning of large and small depends on the criteria used to classify the representative points of the node - reduced representation .
nodes can be classified based on the cd ( relative overlap ) flow , besides distinction based on the sign , the scale is continuous , there is no a - priori grouping of nodes .
further classification can be made based on the structure of the cd ( relative overlap ) flow , i.e. based on properties of different terms defining the cd ( relative overlap ) flow .
components of the flow representation for two toy graphs are shown in figure [ fig : lepke ] .
we observe that same nodes may be global , but not local cd flow sinks or sources . , the right column represents graph nodes with @xmath96 .
every overlapping set is empty for the lower graph , because all chordless circles are of length two .
some points have the same coordinates in the flow representation .
e.g. , point d is is global , but not local cd flow sink . ]
each octant represents different aspect of convergence - divergence relations in the network .
these quantities bring us to the actual interpretation of edge convergence and divergence as a characterisation of signal flow on the nodes of a network . to make statements about the signal flow derived from the cd flow
, we have to make an inversion of properties , as nodes which behave as a sink of convergence , actually inject information to the network , thus they are sources of signal . respectively ,
cd sources are sinks of signal . assuming this interpretation we can extract useful information from the flow representation regarding the signal processing roles of nodes in the network . nodes which have incoming edges with cardinalities of the @xmath11sets ( @xmath12sets ) being larger than cardinalities of the @xmath12sets ( @xmath11sets ) , and outgoing edges with cardinalities of the @xmath12sets ( @xmath11sets ) being larger than cardinalities of the @xmath11sets ( @xmath12sets ) are , from the signal processing perspective , identified as sources of signals .
the combination of divergent input ( negative incoming cd sum ) and convergent output ( positive outgoing cd sum ) is , considering the signal flow , equivalent to absorption of signals in the network .
this is represented in the top left quadrant of the @xmath87 plane . on the opposite , the combination of convergent input and divergent output corresponds to the source characteristics of the nodes ( bottom right quadrant of the @xmath87 plane ) .
the top right and bottom left quadrants can be interpreted as a display of _ signed _ relay characteristics of the nodes .
nodes which have incoming edges with cardinalities of the @xmath12sets ( @xmath11sets ) being larger than cardinalities of the @xmath11sets ( @xmath12ets ) , and outgoing edges with cardinalities of the @xmath12sets ( @xmath11sets)being larger than cardinalities of the @xmath11sets ( @xmath12sets ) , are called negative ( positive ) router nodes .
at the same time routing characteristics can be read from the top right and bottom left quadrants .
redistribute _ incoming cd of a given sign to outgoing cd of the _ same _ sign .
additional information is obtained from the @xmath86 coordinate , which gives the average overlap of incoming and respectively , outgoing edges .
this quantity identifies the degree of a node s participation in signal circulation in the network , a property typically associated with control circuits .
graphical presentation of a network is not unique , e.g. isomorphic graphs may look totally different , the petersen graph being a typical example .
community structure is not unique , grouping of points , thus presenting a network can be achieved in a multitude of ways . yet , the flow representation of a network is _ unique _ , though due to possible symmetries it may have a significant amount of redundancy .
this 3d plot of the network is unique in the sense that there is no arbitrariness in the position of the points in the three dimensional space .
the flow representation can be considered as a network fingerprint since isomorphic graphs are mapped to the same plot , and differences between flow representations can be attributed to structural and functional properties of the network .
if all edges are reciprocal or the graph is undirected , the flow representation of the network shrinks to a single point .
the same argument applies to all graphs in which some nodes can not be distinguished due to symmetries .
more precisely , nodes in the orbit of an element generated by the automorphism group of the graph are represented with the same point on the flow representation , as all the value of @xmath73-s are constants on the orbits generated by the automorphism group of the graph .
usefulness and application of the flow representation will be illustrated in the analysis of the real - world networks in section [ sec : real_net ] .
we calculate cd - s for three model networks and analyse cd - s of four real - world networks . in this section
we analyse functional clusters in real - world networks and the statistical properties of their interconnection .
we analysed two biological and two artificial networks : macaque visuo - tactile cortex @xcite , signal - transduction network of a ca1 neuron @xcite , the call graph of the linux kernel version 2.6.12-rc2 @xcite , and for comparison purposes the street network of rome @xcite .
nodes and edges are defined as follows : in the macaque cortex nodes are cortical areas and edges are cortical fibres , in the signal - transduction network nodes are reactants and edges are chemical reactions , in the call graph nodes are functions and edges are function calls , in the street network the nodes are intersections between roads and edges correspond to roads or road segments .
the first three networks perform computational tasks , linux kernel manages the possibly scarce computational resources , signal - transduction network can be considered as the operating system of a cell , while cortex is an ubiquitous example of a system which simultaneously performs many computationally complex tasks .
the street network is an oriented transportation network , which has a rich structure , as its elements have traffic regulating roles .
the call graph of the linux kernel was constructed in the following way .
we created the call graph of the kernel source which included the smallest number of components necessary to ensure functionality .
the call graph was constructed using the codeviz software @xcite , but it was not identical to the actual network of the functions calling each other , because the software detects only calls that are coded in the source and not the calls only realized during runtime .
the resulting call graph had more than @xmath97 vertices .
as we wanted to perform clustering and statistical tests , the original data was prohibitively large , therefore we applied a community clustering algorithm @xcite to create vertex groups .
we generated a new graph in which the vertices represented the communities of the original call graph and have added edges between vertices representing communities whenever the original nodes in the communities were connected by any number of edges .
definition of the call graph nodes and their connections is analogous to the nodes and connections of the cortical network , as millions of neurons form a cortical area , and two areas are considered to be connected if a relatively small number of neurons in one area is connected to a small number of neurons in another area .
the call graph of the linux kernel will be discussed in section [ sec : aggreg_netw ] .
flow are shown in the left column , components of the average @xmath98 are shown in the right column . displayed are : erds - rnyi graph ( row a ) , macaque visuo - tactile cortex ( row b ) and signal - transduction ( row c ) .
relative overlap flow is indicated by colour intensity . ]
the flow representations of two real - world networks are shown in figure [ fig:3lepke ] and for comparison , in part a , the erds - rnyi network .
we can identify the most important nodes and some general features of the networks as follows .
part b refers to the macaque visuo - tactile cortex .
it is characterised by the alignment of the nodes along a straight line along the main diagonal , and hyperbolic - like pattern in the first and third quadrants , showing reverse ordering in the opposite quadrants , and absence of routers , which refers to a hierarchical organisation . in part c
one can see the signal - transduction network of a hippocampal neuron . in the signal - transduction network of the hippocampal neurons ,
the molecules with the most negative cd flow are involved , among other functions , in the regulation of key participants of the signal transduction cascade such as the camp second messengers .
molecules with large positive cd flow play function in cell survival and differentiation , as well as apoptosis .
router - like proteins are involved in diverse functions , notably the regulation of synaptic transmission in addition to those mentioned above .
however , it should be noted that partly because of the paucity of our knowledge about many of the components of this network , as well as because of redundancy , i.e. overlapping functionality , we could give here only a very superficial classification .
all edges of the signal transduction network fall in one of the three classes : excitatory , inhibitory and neutral , @xcite .
cd and overlap data were unrelated to the inhibitory , excitatory or neutral nature of network edges .
empirical distributions of cd - s and overlaps were alike for each edge class , see figure [ fig : stn_edges ] in the appendix .
we have analysed the flow representations in order to identify different features of signal processing .
network nodes are points represented in a 6d space of the flow representation , and in order to identify different signal processing , transmitting and controlling groups of nodes we performed clustering using gaussian mixture and bayesian information criterion implemented in r @xcite .
we wish to stress that the clustering we performed is not a form of community detection , but grouping of nodes with respect to their functional signal processing properties .
community detection can identify dense substructures , but it provides no information about the nature of signal processing , transmission or control . in each network
we determined local and global , total and average signal processing clusters , have determined their properties , and have analysed the nature of cd - s and relative overlaps within and between clusters .
clustering of nodes with respect to their functional properties resulted in contingency tables , with clusters being labels of the contingency table , and entries in the contingency able being numbers of edges within and between respective clusters . to estimate the randomness of the contingency tables we performed monte carlo implementation of the two sided fisher s exact test .
number of replicates used in the monte carlo test was @xmath97 in each case .
the exact fisher s test characterises the result of the clustering procedure , it quantifies how much the distribution of edges within and between clusters differ .
we summarise the results in table [ tab : eft ] . for comparison purposes benchmark graphs
were generated using algorithms described in @xcite .
.number of functional clusters ( @xmath77 ) and the corresponding @xmath99-values calculated using fisher s exact test of the contingency tables .
q denotes the modularity of the community structure .
two numbers in a single cell denote the first two moments derived from sample size of 100 graph instances .
networks are denoted as follows : vtc - macaque visuo - tactile cortex , stn - signal - transduction network of the hippocampal ca3 neuron , kernel - call - graph of the linux kernel , rome - rome street network , er - erds - rnyi graphs and bench - benchmark graphs .
numbers were rounded to minimise the table size .
definitions of aggregated networks are given in section [ sec : aggreg_netw ] . [ cols="<,^,^,^,^,^,^",options="header " , ] [ tab : eft_2 ] empirical distributions of cd - s and relative overlaps over the excitatory , inhibitory and neutral edge classes in the signal transduction network are shown in figure [ fig : stn_edges ] .
authors are grateful to tams nepusz , lszl ketskemty , lszl zalnyi , balzs ujfalussy , gerg orbn and zoltn somogyvri for useful discussions . 99 amaral lan , scala a , barthlmy m and stanley he 2000 _ proc .
usa _ * 97 * 11149 bnyai m , nepusz t , ngyessy l and bazs f 2009 _ proc .
7-th int . symp . on intelligent systems and informatics ( subotica )
_ 241 ben - naim e and krapivsky pl 2009 _ j. phys . a : math .
theor . _ * 42 * 475001 dorogovtsev sn , mendes jff and samukhin an 2001 _ phys .
e _ * 64 * 025101(r ) erds p and rnyi a 1959 _ publ .
debrecen _ * 6 * 290 http://freshmeat.net/projects/codeviz/ ingram pj , stumpf mph and stark j 2006 _ bmc genom . _ * 7 * 108 jaccard p 1901 _ bull .
de la soc .
nat . _ * 37 * 547 http://kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc2/ lancichinetti a and fortunato s 2009 _ phys . rev .
e _ * 80 * 016118 luo j and magee cl 2011 _ complexity _ doi 10.1002/cplx.20368 maayan a , jenkins sl , neves s , hasseldine a , grace e , dubin - thaler b , eungdamrong nj , weng g , ram pt , rice jj , kershenbaum a , stolovitzky ga , blitzer rd and iyengar r 2005 _ science _ * 309 * 1078 ngyessy l , nepusz t , kocsis l and bazs f 2006 _ eur . j. neurosci . _
* 23 * 1919 ngyessy l , nepusz t , zalnyi l and bazs f 2008 _ proc .
b _ * 275 * 2403 newman mej , strogatz sh and watts dj 2001 _ phys
* 64 * 026118 newman mej 2003 _ siam rev . _
* 45 * 167 pons p and latapy m 2005 _ computer and information sciences - iscis 2005 ( istanbul ) _ vol .
( berlin / heidelberg : springer ) 284 http://r-project.org http://www.dis.uniroma1.it/@xmath100challenge9/data/rome/rome99.gr stelling j , klamt s , bettenbrock k , schuster s and gilles ed 2002 _ nature _ * 420 * 190 sporns o , 2002 pp .
171 - 186 , graph theory methods for the analysis of neural connectivity patterns , ed r ktter _ neuroscience databases
. a practical guide _
( boston , ma : kluwer ) | confining an answer to the question whether and how the coherent operation of network elements is determined by the the network structure is the topic of our work .
we map the structure of signal flow in directed networks by analysing the degree of edge convergence and the overlap between the in- and output sets of an edge .
definitions of convergence degree and overlap are based on the shortest paths , thus they encapsulate global network properties . using the defining notions of convergence degree and overlapping set we clarify the meaning of network causality and demonstrate the crucial role of chordless circles . in
real - world networks the flow representation distinguishes nodes according to their signal transmitting , processing and control properties .
the analysis of real - world networks in terms of flow representation was in accordance with the known functional properties of the network nodes .
it is shown that nodes with different signal processing , transmitting and control properties are randomly connected at the global scale , while local connectivity patterns depart from randomness .
grouping network nodes according to their signal flow properties was unrelated to the network s community structure .
we present evidence that signal flow properties of small - world - like , real - world networks can not be reconstructed by algorithms used to generate small - world networks .
convergence degree values were calculated for regular oriented trees , and its probability density function for networks grown with the preferential attachment mechanism . for erds - rnyi graphs we calculated both the probability density function of convergence degrees and of overlaps . |
when considering the dispersion of dielectric media , text book analysis typically concerns itself with sums of lorentzian functions @xcite . while it can be argued that sums of lorentzians are physically reasonable response functions for a number of systems , in particular those described by the lorentz - model @xcite , it is well known that the only restrictions imposed by causality are those implied by the kramers - kronig relations @xcite . in light of the variety in the electromagnetic responses offered by natural media and metamaterials ,
it is of interest to consider the possible gap between the function space consisting of sums of lorentzians , and the space consisting of functions satisfying the kramers - kronig relations . to this end
, it is here demonstrated that any complex - valued function satisfying the kramers - kronig relations can be approximated as a superposition of lorentzian functions , to any desired accuracy .
it therefore follows that the typical analysis of causal behavior in terms of lorentzian functions for dielectric or magnetic media encompasses the whole space of functions obeying the kramers - kronig relations .
these results therefore serve to strengthen the generality of the typical analysis of causality .
two examples where the response functions do not resemble typical lorentzian resonance behavior shall here be expressed as lorentzian superpositions in order to demonstrate the above findings .
section [ sec : nies ] considers a steep response function which results in a susceptibility @xmath0 for arbitrarily low loss or gain @xcite , and sec .
[ sec : perfectlense ] considers an optimal perfect lens response over a bandwidth @xcite .
the precisions in both superpositions are shown to become arbitrarily accurate as the parameters are chosen appropriately .
a natural consequence of such superpositioning is that lorentzian functions can be viewed as general building blocks for engineering causal susceptibilities in metamaterials .
considering that systems such as the pioneering split - ring resonator implementation @xcite , and others @xcite , have demonstrated several ways of realizing and tailoring lorentzian responses , this may prove to be a promising approach .
while literature has so far tended to focus on specific metamaterial designs , a number of desired responses have emerged for which few physically viable systems are known @xcite .
one such set of response functions are those with desired dispersion properties @xcite which are relevant for applications such as dispersion compensation @xcite , couplers @xcite , antenna design @xcite , filters @xcite , broadband absorption @xcite and broadband ultra - low refractive index media @xcite .
this leads to the following question : starting with a target response , how can one realize an approximation of it ? towards this end , it has been proposed to use layered metamaterials @xcite .
our article considers more generally the possibility of engineering artificial response functions through the realization of a finite number of lorentzians ; a method which may be applicable to a variety of metamaterials . on this note sec .
[ sec : tailorsplitring ] addresses how lorentzian superposition responses can be realized through the arrangement of split ring cylinders of different radii and material parameters .
finally sec .
[ sec : errorestimate ] derives an estimate of the error that arises when a target response is approximated by a finite sum of lorentzian functions .
the following section will set out the main results of this article while leaving detailed calculations to later sections and appendices .
a lorentzian function can be written in the form @xmath1 where @xmath2 is the resonance frequency , @xmath3 is the frequency , and @xmath4 is the bandwidth
. it may be demonstrated that the imaginary part of @xmath5 approaches a sum of two dirac delta - functions with odd symmetry as @xmath6 when @xmath7 .
this is exemplified in fig .
[ fig : lorconvergence ] and proven in appendix [ sec : deltaconv ] . a goal function , such as the imaginary part of a susceptibility @xmath8 , may therefore be expressed @xmath9 this limit integral expression may then be approximated by the sum approaching a sum of two delta - functions for decreasing values of @xmath10.,scaledwidth=44.0% ] @xmath11 where @xmath12 , @xmath13 is the resolution along the integration variable @xmath14 , and @xmath15 is a large integer .
this sum , which shall be designated @xmath16 , can be made to approximate @xmath8 to any desired degree of accuracy .
more precisely , the involved error is shown to converge to zero in both @xmath17 and @xmath18 : @xmath19 where @xmath13 is chosen suitably , e.g. @xmath20 ( appendix [ sec : l2conv ] ) .
considering , one may define a function @xmath21 where now both real and imaginary parts of the lorentzians are superposed . from the kramers - kronig relations
one then has @xmath22 where @xmath23 represents the hilbert transform and @xmath24 the real part of . since the hilbert transform preserves the energy , or @xmath18-norm , it therefore follows that @xmath25 when given . on the basis of this
, it follows that @xmath26 meaning that both real and imaginary parts of @xmath27 are approximated by the summation of lorentzians .
the terms are weighted by @xmath28 at each resonance frequency according to .
. combined with therefore becomes the central result of this article .
noting that is itself a riemann sum , it follows that the limit integral expression @xmath29 may be written on the basis of the preceding arguments .
therefore approximate the space of functions satisfying the kramers - kronig relations as superpositions of lorentzians to any degree of precision . in the following section
this result shall be demonstrated with two examples .
the validity of requires that @xmath27 is analytic on the real axis ( not only in the upper half - plane ) . in the event of non - analytic susceptibilities on the real axis , however , all problems are bypassed by instead evaluating @xmath8 along the line @xmath30 before approximating by lorentzians .
here @xmath31 is an arbitrarily small parameter . since @xmath27 is analytic there , is valid .
furthermore , since @xmath32 almost everywhere as @xmath33 , the representation can be made arbitrarily accurate , meaning that any @xmath8 can be approximated to any precision .
in fact , this also includes media that do not strictly obey the kramers - kronig relationship due to singularities on the real axis , such as the ideal plasma .
it is possible to achieve @xmath34 with arbitrarily low loss or gain @xcite .
consider a susceptibility with @xmath35 as a result of the infinite steepness at @xmath36 , the hilbert transform gives @xmath37 .
it follows that it is possible to scale to make the magnitude @xmath38 arbitrarily small for all frequencies while maintaining @xmath39 .
inserting as @xmath40 in gives @xmath41 .
\label{eq : niesint}\end{aligned}\ ] ] one can show that as @xmath7 , the imaginary part of yields . using one may likewise approximate the response as a finite sum of lorentzians : fig .
[ fig : nies ] plots the real and imaginary parts of @xmath27 as approximated by both sum and integral expressions , and respectively , where @xmath42 is chosen equal to @xmath43 , @xmath44 , and @xmath45 , and where @xmath46 .
one observes that the sum falls in line with the integral result . for @xmath47 one
has @xmath48 . as @xmath49 , meaning that the drop in the approximated curves of @xmath8 at @xmath50 become infinitely steep
, one has that @xmath51 in both cases .
the imaginary parts for one out of every 5 lorentzians in the sum are displayed in fig .
[ fig : summedlor ] for @xmath52 .
+ it has been shown in @xcite that the resolution , @xmath53 , for a metamaterial lens of thickness @xmath54 surrounded by vacuum , is given by @xmath55 for a one dimensional image object . here , @xmath27 is either the electric or magnetic susceptibility depending on the polarization of the incident field .
it follows that any perfect lens should approximate @xmath56 over a bandwidth .
a system approaching such an optimum is displayed in fig .
[ fig : perflens ] @xcite .
there has been placed a strong lorentzian resonance at @xmath57 ( out of view ) and a weak , slowly varying function around @xmath58 .
taking the absolute value gives the black dashed curve in fig .
[ fig : absperflenssum ] , which reveals that @xmath59 remains small and constant over the bandwidth .
+ + + fig .
[ fig : perflenssum ] represents a sum of lorentzians over the interval @xmath60 where @xmath61 and @xmath62
. the goal function for has been found by subtracting the strong resonance situated at @xmath57 from @xmath8 in fig .
[ fig : perflens ] .
the strong resonance has then later been added to the sum of lorentzians .
the absolute value of the real and imaginary parts in fig .
[ fig : perflenssum ] correspond with the green solid curve in fig .
[ fig : absperflenssum ] , which neatly follows the dashed curve of @xmath59 . also observed in fig .
[ fig : absperflenssum ] is the corresponding result of another sum approximation where the lorentzians are wider .
hence one observes the trend that as @xmath10 decreases the lorentzian sum approximates @xmath59 increasingly well .
[ fig : perflenssumcomp ] displays the imaginary parts for one out of every five lorentzians in the sum corresponding with fig . [
fig : perflenssum ] .
sections [ sec : nies ] and [ sec : perfectlense ] have demonstrated that useful responses which do not arise in any natural systems can nevertheless be approximated as a superposition of ordinary lorentzian resonances to any desired precision . considering that lorentzian resonances can be realized and tailored for both permittivities @xmath63 and permeabilities @xmath64 in numerous metamaterial realizations @xcite
, this is of particular interest for the prospect of engineering desired metamaterial responses .
for instance , in the case of an array of split ring cylinders one may find for the effective permeability @xcite : @xmath65 where @xmath66 is the capacitance per @xmath67 in the split ring cylinder , @xmath53 is its radius , @xmath68 is the resistance per circumference - length ratio and @xmath69 is the fractional volume of the unit cell occupied by the interior of the cylinder . for small bandwidths
the presence of @xmath70 in the numerator is not significant and approximates a lorentzian response . by comparing it with one
may determine the lorentzian parameters as @xmath71 the resonator strength @xmath69 is expressed @xmath72 where @xmath73 is the dimension of the unit cell .
hence the resonance frequencies , widths , and strengths can be tailored by varying @xmath53 , @xmath68 , @xmath66 .
it may be noted that split ring cylinders display large resistivity in the optical regime , making it difficult to achieve narrow lorentzian responses there .
for this reason other systems , such as metamolecules of nanoparticles , have been proposed for optical purposes @xcite . in order to realize a sum of lorentz resonances such as those leading to fig .
[ fig : summedlor ] and fig .
[ fig : perflenssumcomp ] by means of split ring cylinders , one could propose to place different cylinders in each unit cell .
the idea would be to realize and superpose a number of different lorentzian resonances of the form corresponding with different values of @xmath53 , @xmath68 , @xmath66 .
the density of each type of split ring cylinder can then be thought to give the appropriate resonance strength .
however , as it is known that two dipoles in close vicinity influence each others dipole moment , it is not intuitively clear that the total response will be as simple as a superposition of the individual responses . the remainder of this section will therefore be used to demonstrate this . towards this end , the procedure outlined in @xcite will be modified to calculate the effective permeability @xmath74 of a split ring cylinder metamaterial when every unit cell contains two split ring cylinders of different radii ( fig .
[ fig : unitcell ] ) .
the effective permittivity @xmath74 is expressed by finding the effective macroscopic fields @xmath75 and @xmath76 over the array from the corresponding actual fields @xmath77 and @xmath78 in each unit cell : @xmath79 one may show that @xmath80 where @xmath81 is the surface current per unit length along the circumference of the cylinders with radii @xmath82 and @xmath83 respectively . when finding an expression for the field inside each cylinder @xmath84 and @xmath85 , is used to express @xmath86
since both split ring cylinders observe the same interaction field contained in @xmath76 , application of faraday s law to each cylinder leads to two equations that may be solved individually for @xmath87 and @xmath88 : @xmath89 in order to evaluate , one may now rearrange to find an expression for @xmath90 in terms of @xmath87 and @xmath88 , for which one in turn can use to find @xmath91 here @xmath92 and @xmath93 is the volume fraction occupied by each split ring cylinder . by comparison with
one observes that the resulting @xmath94 here is simply the superposition of the individual split ring cylinder responses as found in @xcite .
this comes as a consequence of the interaction field sensed by both split ring cylinders being uniform .
this description of the interaction is valid as long as the cylinders are long , for which the returning field lines at the end of each cylinder are spread uniformly over the unit cell . finally , the analysis presented here is easily generalized for unit cells with many different split ring cylinders , for which remains valid and @xmath95 labels the different split ring cylinders in the unit cell .
( 1,1.11714446 ) ( 0,0 ) containing two different cylinders.,title="fig : " ] ( 0.18851057,0.95124317)(0,0)[lb ] ( 0.53739076,0.72547227)(0,0)[lb ] ( 0.48,0.00665475)(0,0)[lb ] towards the goal of engineering a desired response @xmath27 by use of a finite number of lorentzians with non - zero widths ( @xmath96 ) , it is of interest to quantify the precision of such an approximation .
an error estimate can be found by considering the combined error introduced in both the delta - function approximation of @xmath8 for finite @xmath10 , and the riemann sum approximation of the integral .
the total error may therefore be bounded according to @xmath97 where @xmath98 represents without the limit and @xmath16 represents the sum . in what follows , bounds for the riemann sum approximation error @xmath99 and delta - convergence error @xmath100 shall be derived in turn .
defining @xmath101 , and naming the integrand in @xmath102 , one may find an upper bound on the riemann approximation error to be @xmath103 where @xmath104 in obtaining the last inequality in , it has been used that @xmath105 is dominated by the curvature of the lorentzian for sufficiently small @xmath10 . after some algebra one finds @xmath106 when assuming that @xmath107 .
note that can be rewritten @xmath108 { \mathrm{im}}\chi ( \omega_0 ) \nonumber \\ & & { \mathrm{im}}\bigg \ { \frac{\omega_0}{\omega_0 ^ 2-\omega^2-i\omega\gamma } \bigg \ } { \text{d}}\omega_0.\end{aligned}\ ] ] by taking the limit @xmath49
this expression becomes of the form , and hence the definition of delta convergence can be used to find the limit of @xmath109 : @xmath110 { \mathrm{im}}\chi(\omega).\end{aligned}\ ] ] it is observed that @xmath111 vanishes as @xmath112 .
for @xmath113 the error involved in replacing @xmath109 by its limiting value becomes small .
therefore , for most goal functions @xmath8 , where the sum covers the frequency interval of interest , the error described by @xmath111 is roughly as small as @xmath38 outside of this interval .
the delta - convergence approximation error @xmath114 which arises when using lorentzians of finite width ( @xmath115 ) rather than delta - functions in , shall now be evaluated .
consider fig .
[ fig : deltakonv ] : the imaginary part of an arbitrary response @xmath27 is to be approximated according to with finite @xmath10 , and the positive frequency peak of the imaginary part of a lorentzian @xmath116 is displayed .
the lorentzian can be expanded in terms of partial fractions as : @xmath117 where @xmath118 corresponds to positive and negative frequency peaks of the lorentzian respectively , according to : ( 1,0.52763968 ) ( 0,0 ) to be expressed as a superposition of finite - sized lorentzians.,title="fig : " ] ( 0.480,0.077)(0,0)[lb ] ( 0.95,0.06913388)(0,0)[lb ] ( 0.480,0.00977939)(0,0)[lb ] ( 0.167009,0.43)(0,0)[lb ] ( 0.52,0.27648538)(0,0)[lb ] @xmath119 introducing into ( without the limit ) then allows the delta - convergence error @xmath120 to be expressed as @xmath121 here , by having made an appropriate substitution , it is observed in the last line that the integral can be written in terms of the lorentz - cauchy function @xmath122 : @xmath123 as a consequence of the substitution , a new function @xmath124 is defined in which essentially represents a shifted version of @xmath40 .
the integration interval in may be divided into three intervals corresponding to the intervals designated in fig . [
fig : deltakonv ] , which shall then be evaluated separately : @xmath125 here the parameter @xmath73 is in principle arbitrary , but can be thought to represent the region around @xmath126 where @xmath127 is assumed to be taylor expandable in the lowest orders .
evaluating integrals ( a ) and ( c ) in together , an upper bound on their sum may be found : @xmath128 \label{eq : upperboundside}\end{aligned}\ ] ] here it has been used that @xmath129 .
one observes that as @xmath130 the @xmath131-term can be expanded , allowing to be re - expressed as @xmath132 which evidently converges to zero when @xmath49 as long as @xmath28 is bounded . in order to calculate ( b )
, @xmath124 is expanded around @xmath133 : @xmath134 \label{eq : taylorexpnded } \ ] ] due to the parity when inserting this into ( b ) only even order terms remain , giving @xmath135 only @xmath10-terms to the first order have been kept in arriving at this expression .
the term @xmath136 corresponds to terms containing higher order derivatives of @xmath127 , which will be assumed to be negligible . in the event
that there exists significant higher order derivatives ( e.g. as will be the case in fig .
[ fig : example1 ] near the steep drop ) , one may keep more terms in going from to until the next even ordered derivative is negligible .
the same steps that now follow may then be applied in order to find the relevant error estimate . in order to arrive at an expression for @xmath120 in terms of the goal function @xmath8 and its derivatives ,
the shifted function @xmath137 is expanded under the assumption that @xmath107 and then inserted for @xmath127 and its second derivative in .
expanding @xmath131 under the assumption that @xmath130 permits further simplification . combining the resulting expression with , yields an upper bound on the delta - convergence error @xmath138 where @xmath139 arises after having expanded and replaced @xmath127 and its second derivative . minimizing the upper bound with respect to @xmath73 then gives @xmath140 an inverse relationship between @xmath141 and @xmath142 is intuitive given that @xmath73 must be small when @xmath8 varies steeply .
inserting into while neglecting @xmath139 gives @xmath143 hence , the delta - convergence error @xmath120 is proportional to @xmath144 under the assumption that @xmath145 , the higher order derivatives are negligible , and @xmath107 .
[ fig : perflenserrorbound ] shows applied to the imaginary part of the perfect lens response displayed in fig .
[ fig : perflenssum ] , where one observes that the actual error ( black curve ) is bounded by ( dashed curve ) .
the displayed error is actually taken to be the total error @xmath146 of the superposition , rather than the difference between @xmath8 and @xmath147 .
however , here @xmath148 in the sum has been reduced by a factor 2 as compared to in fig .
[ fig : perflenssum ] , in order to reduce the significance of the riemann sum error in comparison to the delta - convergence approximation error . with a lorentzian superposition
is shown to be bounded by .,scaledwidth=44.0% ]
it has been demonstrated that superpositions of lorentzian functions are capable of approximating all complex - valued functions satisfying the kramers - kronig relations , to any desired precision .
the typical text - book analysis of dielectric or magnetic dispersion response functions in terms of lorentzians is thereby extended to cover the whole class of causal functions .
the discussion started with approximating the imaginary part of an arbitrary susceptiblity @xmath27 by a superposition of imaginary parts of lorentzian functions .
the error @xmath149 was then shown to become arbitrarily small as @xmath49 , @xmath112 and @xmath150 . since
this error was shown to vanish in @xmath18 it was known that @xmath151 . from this
the superposition @xmath152 was found , which approximates @xmath27 to any desired precision .
the superposition @xmath152 has been demonstrated to reconstruct two response functions that obey the kramers - kronig relations while not resembling typical lorentzian resonance behavior .
the first response is known to lead to significant negative real values of the susceptibility @xmath27 in a narrow bandwidth with arbitrarily low gain or loss , while the other response represents the optimum realization of a perfect lens on a bandwidth .
these examples demonstrate the possibility of viewing lorentzians as useful building - blocks for manufacturing desired responses that are not found in conventional materials . to this end
, the ability to realize lorentzian sums using split ring cylinders with varying parameters was shown . in order to quantify the precision of a superposition approximation ,
error estimates have been derived .
this section will prove and equivalently . in doing so , observe that an implicit definition of the delta - function is @xmath153 { \text{d}}\omega_0 .
\nonumber \\\end{aligned}\ ] ] note that the last line holds only insofar as the goal function obeys odd symmetry @xmath154 . by inserting into and using the symmetry @xmath155
, one observes that the goal of this section becomes to demonstrate that @xmath156 inserting for @xmath116 in the integral and making the appropriate substitution leads to : @xmath157 one may then move the limit inside the integral under the criteria of lebesgue s dominated convergence theorem . as @xmath49 the function @xmath158 converges pointwise to @xmath8 .
furthermore , it is clear that an integrable function that bounds the integrand for all values of @xmath10 exists under the condition that @xmath8 is bounded .
hence , gives @xmath159 which concludes the proof .
this section considers the conditions upon @xmath10 , @xmath13 and @xmath160 needed in order to obtain the @xmath18 convergence expressed in . from the discussion in sec .
[ sec : errorestimate ] , one may express : the riemann sum error is expressed through and , and the delta - convergence error from and is expressed in , and .
the term @xmath139 in is included due to having expanded and replaced @xmath127 and its second derivative so as to express in terms of @xmath8 and its second derivative .
the bound given by - has been derived under the assumption that @xmath162 . before proceeding to evaluate the @xmath18 norm by these
, one must therefore demonstrate that @xmath163 is bounded in a region @xmath164 where @xmath165 ( representing the region not accounted for by - ) , so that as @xmath49 and then @xmath166 , it is known that the @xmath18 norm converges to zero also here .
a detailed proof of this is will not be given here , but its result can be understood by noting that , when without its limit , is bounded by an integral of @xmath167 on the interval @xmath168 multiplied with @xmath169 . since the integral is finite and @xmath27 is analytic , it follows that is bounded . furthermore , since the sum can be made arbitrarily close to , it is intuitive that it also remains bounded for all @xmath3 . since
@xmath8 and @xmath16 are bounded it follows that @xmath163 is finite for all @xmath3 .
proceeding now to evaluate the @xmath18 norm from - , convergence for is achieved by setting @xmath170 and taking the limit @xmath49 , since @xmath8 is in @xmath18 ( the presence of the term @xmath171 will be discussed below ) .
the same convergence occurs in as @xmath49 .
the @xmath18 norm arising for converges provided that @xmath173 is in @xmath18 .
since it however is conceivable for a function @xmath27 in @xmath18 to have derivatives not in @xmath18 , one may instead consider @xmath27 along the line @xmath174 , where @xmath175 is an arbitrarily small parameter .
one then finds through the fourier transform @xmath176 \text{e}^{i\omega t } { \text{d}}t , \label{eq : fourierintegral}\end{aligned}\ ] ] given that @xmath177 is the time - domain response associated with @xmath27 .
reveals that one may express @xmath178 since @xmath177 is in @xmath18 one observes that @xmath179 is in @xmath18 for @xmath175 . knowing that the fourier transform preserves the @xmath18 norm
, it follows that @xmath180 is in @xmath18 . the terms @xmath171 and @xmath139 in and involve multiples of either @xmath8 or its derivatives . from it
is clear that derivatives of all orders are assured to be in @xmath18 by the above procedure .
hence these terms vanish as @xmath49 .
note that in the derivation of @xmath181 , derivatives greater than the second order of @xmath8 have been neglected , as discussed with regards to .
if these were to be included here , they would lead to terms of the same form as and would converge in the same manner .
considering now , one observes that by using one may express this as : @xmath182 to find the limit of the @xmath18 norm , the task here is therefore to evaluate @xmath183moving the limit inside the outermost integral is permitted through lebesgue s dominated convergence theorem , under the condition that @xmath184 is bounded . in the event that @xmath184 is singular on the real axis , one may instead use @xmath185 ( where @xmath31 is an arbitrarily small parameter ) for which any divergence is quenched , as discussed in the introduction .
this gives @xmath186 when having used .
the final result reveals that the norm is finite , and that one however must demand @xmath112 in order that converges to zero .
the remaining task is to evaluate through the limit @xmath187 under the same condition as before one may move the limit inside the outermost integral through lebesgue s dominated convergence theorem : @xmath188 here the delta - convergence property of the lorentz - cauchy function @xmath122 has been used .
20ifxundefined [ 1 ] ifx#1 ifnum
[ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _ _ ( , ) http://books.google.no/books?id=tyeakaeacaaj[__ ] ( , ) http://books.google.no/books?id=wv-wrtrdwt0c[__ ] ( , ) link:\doibase 10.1109/jproc.2011.2114631 [ * * , ( ) ] @noop _ _ ( , ) @noop * * , ( ) link:\doibase 10.1364/josab.30.000370 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.99.107401 [ * * , ( ) ] link:\doibase 10.1109/tmtt.2003.819193 [ * * , ( ) ] link:\doibase 10.1109/tmtt.2005.845188 [ * * , ( ) ] link:\doibase 10.1109/tmtt.2004.823579 [ * * , ( ) ] link:\doibase 10.1109/tap.1962.1137809 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1364/ol.37.002133 [ * * , ( ) ] link:\doibase 10.1364/josab.20.002448 [ * * , ( ) ] link:\doibase 10.1063/1.4746400 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) | we prove that all functions obeying the kramers - kronig relations can be approximated as superpositions of lorentzian functions , to any precision . as a result
, the typical text - book analysis of dielectric dispersion response functions in terms of lorentzians may be viewed as encompassing the whole class of causal functions .
a further consequence is that lorentzian resonances may be viewed as possible building blocks for engineering any desired metamaterial response , for example by use of split ring resonators of different parameters .
two example functions , far from typical lorentzian resonance behavior , are expressed in terms of lorentzian superpositions : a steep dispersion medium that achieves large negative susceptibility with arbitrarily low loss / gain , and an optimal realization of a perfect lens over a bandwidth .
error bounds are derived for the approximation . |
high energy gamma ray emission from the crab pulsar was observed by satellite - borne telescopes for 15 years : in 1972 - 73 by sas-2 ( kniffen et al .
1974 ) , from 1975 - 82 by cos - b ( clear et al . 1987 ) , and since 1991 by egret ( nolan et al . 1993 ; ramanamurthy et al . 1995 ) .
early observations showed possible sinusoidal variation in the relative intensities of the two peaks ( wills et al .
1982 ) with a time scale of @xmath0 years , and it was suggested that this variation might be due to the precession or free nutation of the neutron star ( kanbach 1990 , zel 1991 ) .
an apparent confirmation of the sinusoidal signal was seen in the low energy gamma ray emission ( ulmer et al .
1994 ) , matched in phase and period with the high energy results , but with a smaller amplitude .
egret data from 1991 through early 1994 were consistent with the expected variation ( nolan et al .
1993 , ramanamurthy et al . 1995 ) , although these observations spanned a time when the ratio of the peaks was predicted to be fairly constant , near the minimum of the sinusoid .
egret observations have now extended the available data by over two years .
the most recent data were expected to be @xmath1 from the average of the previous values if the sinusoidal model is correct .
the egret instrument is a spark chamber gamma - ray telescope with an energy range of 30 mev 30 gev .
details of the instrument design , calibration , and standard analysis software are given in thompson et al .
( 1993 ) .
all viewing periods where egret was pointed within @xmath2 of the crab were analyzed , with the exception of viewing period 0021 ( 1991 jul 815 ) , in which there was a large solar flare .
the compton observatory viewing period numbers and dates for these observations are shown in table 1 .
the eight viewings numbered 4120 through 5280 ( 1995 feb . through 1996 aug . )
have been completed since the time of the previous work of ramanamurthy et al .
( 1995 ) .
all photons with measured energy above 50 mev , which were within an energy - dependent cone of half - angle @xmath3 were used in this analysis .
the angle @xmath3 , chosen such that @xmath4 of the photons originating from the pulsar are within the acceptance cone , is given by ( thompson et al .
1993 ) @xmath5 the arrival time of each detected photon was transformed to solar system barycentric time using the de200 ephemeris , then binned according to the pulsar phase at that time , determined from the princeton pulsar timing database ( arzoumanian et al .
this analysis was performed using the pulsar program ( fierro 1995 ) . as seen in figure 1 ,
the light curve was divided into several sections , including peak 1 ( phase .94 .04 ) , peak 2 ( phase .32 .46 ) , and the off - pulse background ( phase .46 .94 ) .
these definitions follow those used in the cos - b analysis ( wills et al .
1982 ) and are similar to those used by nolan et al .
( 1993 ) and ramanamurthy et al .
the background ( from the crab nebula , nearby sources , and the diffuse galactic radiation ) was assumed constant as a function of pulsar phase .
the off - pulse count rate was then used to find the background - subtracted counts estimates of the two peaks ( @xmath6 and @xmath7 ) . in order to avoid any effects of changes in instrument performance ,
the evolution of the ratio @xmath8 was examined ( as in previous analyses ) .
the two peaks have slightly different energy spectra ( nolan et al .
1993 ) , and the egret response at different energies has changed at different rates ( esposito et al .
thus the ratio @xmath8 is affected by the changes in instrument performance over time .
calculations of this effect , however , indicate that it is an order of magnitude smaller than the errors in @xmath8 due to poisson fluctuations .
the differences in the energy responses of the sas-2 , cos - b , and egret instruments are larger than the variation in the egret response . however , statistical errors in the previous instruments data are larger as well .
thus the value of @xmath8 obtained from egret data should be comparable to that obtained with sas-2 and cos - b .
the values of @xmath8 obtained , together with the @xmath9 errors , are shown in table 1 .
a reduced data set , where nearby points are joined for clarity , is shown in figure 2 .
the data were fit with a constant , yielding @xmath10 , with @xmath11 with 20 degrees of freedom ( dof ) .
such a low value of @xmath12 might imply that the errors in the data were over - estimated . in this case , however , the errors arise purely from statistical poisson fluctuations , and the low value must occur purely by chance .
the data are very consistent with a constant value of @xmath8 .
the data were also fit with a straight line ( @xmath13 with 19 dof ) , and a quadratic ( @xmath14 with 18 dof ) .
neither result gives a significantly better fit : the egret data are most consistent with no variation .
the egret data were also analyzed in conjunction with the sas-2 ( kanbach 1990 ) and cos - b ( clear et al .
1987 ) data .
the best fit sinusoid to the previous instruments data , @xmath15 where t is the year of the observation , and to the combined data set , @xmath16 are shown together with the data and the average value of p2/p1 in figure 3 . as can be seen ,
the most recent egret observations ( the last data point ) are much less consistent with the large amplitude sinusoid .
the constant value of @xmath17 gives a @xmath18 ( with 27 dof ) , indicating a good fit .
the sinusoid fit to the combined data gives a period of 11.6 years with @xmath19 ( with 24 dof ) , which does not represent a significant improvement .
thus , using the combined data sets , the data are most consistent with no variation in p2/p1 .
the light curves obtained from phase 1 data ( apr 91sept 92 ) and from phases 4 and 5 ( feb 95aug 96 ) are shown in figure 1 .
the overall shape seems to have changed little , in contrast with the expected change if the 14 year cycle were correct .
a @xmath12 test was performed to compare the two light curves in a quantitative way . in order to take out possible systematic effects , both an additive offset and a multiplicative factor for the second light curve
were fit to the data .
the resulting @xmath20 ( with 48 dof ) is consistent with no change in the light curve . without the offset , @xmath21 ( with 49 dof )
was obtained , indicating an inconsistency at the 99.95% confidence level .
the offset required indicates a lower background level in the later observations .
this could be due to a change in the nebular emission ( de jager et al .
1996 ) , or could be a result of changes in the performance of egret . as the gas in the spark chamber ages , the sensitive area at low energies decreases , which decreases the width of the average point spread function .
this effect might lower the background in the later observations .
recent observations with egret have provided data relevant to the reported variation in the p2/p1 ratio of the crab pulsar .
data from sas-2 and cos - b suggested a sinusoidal variation in this ratio .
the egret data , both taken alone and in conjunction with the data from previous instruments , are most consistent with a constant value of p2/p1 .
examination of the light curves from early and later observations shows no distinct changes in the pulsar s light curve .
the egret data can not , of course , rule out past variability in the p2/p1 ratio .
future observations by egret and its successors may allow a more precise characterization of the long term behavior of the crab s light curve .
the egret team gratefully acknowledges support from the following : the arcs foundation ( wft ) , bundesministerium fr bildung , wissenschaft , forschung und technologie grant 50 qv 9095 ( mpe ) , nasa cooperative agreement ncc 5 - 95 ( hsc ) , nasa grant nag5 - 1605(su ) , and nasa contract nas5 - 96051 ( ngc ) .
lrl 0002 & apr 2228 1991 & @xmath22 0003 & apr 28may 1 1991 & @xmath23 0004 & may 14 1991 & @xmath24 0005 & may 47 1991 & @xmath25 0010 & may 1630 1991 & @xmath26 0360 & aug 1112 1992 & @xmath27 0365 & aug 1220 1992 & @xmath28 0390 & sep 117 1992 & @xmath29 2130 & may 2329 1993 & @xmath30 2210 & may 1324 1993 & @xmath31 3100 & dec 113 1993 & @xmath32 3211 & feb 815 1994 & @xmath33 3215 & feb 1517 1994 & @xmath34 4120 & feb 28mar 7 1995 & @xmath35 4130 & mar 721 1995 & @xmath36 4200 & may 23jun 6 1995 & @xmath37 4260 & aug 822 1995 & @xmath38 5020 & oct 1731 1995 & @xmath39 5260 & jul 30aug 13 1996 & @xmath40 5270 & aug 1320 1996 & @xmath41 5280 & aug 2027 1996 & @xmath42 arzoumanian .
z. , nice , d. , & taylor , j.h .
1992 , gro / radio timing data base , princeton university clear , j. et al .
1987 , , 174 , 85 esposito , j.a .
1997 , in preparation .
de jager , o.c .
1996 , , 457 , 253 fierro , j. m. 1995 , ph.d .
thesis , stanford university kanbach , g. 1990 , in the egret scinece symposium : nasa conference publication no .
fichtel , c.e .
101 kniffen , d.a .
1974 , 251 397 nolan , p.l .
1993 , , 409 , 697 zel , m.e .
1991 , europhysics letters , 14 , 3 ramanamurthy , p.v .
et al , 1995 , , 450 , 791 thompson , d.j .
1993 , , 86 , 629 ulmer , m.p .
1994 , , 432 , 228 wills , r.d .
1982 , nature , 296 , 723 | recent observations of the crab pulsar by the energetic gamma - ray experiment telescope ( egret ) on the compton gamma ray observatory show that the high - energy gamma - ray light curve has changed little over the lifetime of the instrument .
previous data collected by sas-2 and cos - b in the years 1972 - 82 , along with earlier egret data , suggested a 14 year sinusoidal variation in the flux ratio between the first and second peaks .
the new data from egret indicate that the flux ratio is constant .
accepted astrophysical journal . scheduled 20 sep 1997 |
statistical models can be broadly classified into _
parametric _ and _ nonparametric _ models .
parametric models , indexed by a finite dimensional set of parameters , are focused , easy to analyse and have the big advantage that when correctly specified , they will be very efficient and powerful .
however , they can be sensitive to misspecifications and even mild deviations of the data from the assumed parametric model can lead to unreliabilities of inference procedures .
nonparametric models , on the other hand , do not rely on data belonging to any particular family of distributions . as they make fewer assumptions ,
their applicability is much wider than that of corresponding parametric methods .
however , this generally comes at the cost of reduced efficiency compared to parametric models .
standard time series literature is dominated by parametric models like autoregressive integrated moving average models @xcite , the more recent autoregressive conditional heteroskedasticity models for time - varying volatility @xcite , state - space @xcite , and markov switching models @xcite . in particular , bayesian time series analysis @xcite is inherently parametric in that a completely specified likelihood function is needed .
nonetheless , the use of nonparametric techniques has a long tradition in time series analysis .
@xcite introduced the periodogram which may be regarded as the origin of spectral analysis and a classical nonparametric tool for time series analysis @xcite .
_ frequentist _ time series analyses especially use nonparametric methods @xcite including a variety of bootstrap methods , computer - intensive resampling techniques initially introduced for independent data , that have been taylored to and specifically developed for time series @xcite . an important class of nonparametric methods is based on frequency domain techniques , most prominently smoothing the periodogram .
these include a variety of frequency domain bootstrap methods strongly related to the whittle likelihood @xcite and found important applications in a variety of disciplines @xcite .
despite the fact that _ nonparametric bayesian _ inference has been rapidly expanding over the last decade , as reviewed by @xcite , @xcite , and @xcite , only very few nonparametric bayesian approaches to time series analysis have been developed .
most notably , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite used the whittle likelihood @xcite for bayesian modelling of the spectral density as the main nonparametric characteristic of stationary time series .
the whittle likelihood is an approximation of the true likelihood .
it is exact only for gaussian white noise .
however , even for non - gaussian stationary time series which are not completely specified by their first and second - order structure , the whittle likelihood results in asymptotically correct statistical inference in many situations .
as shown in @xcite , the loss of efficiency of the nonparametric approach using the whittle likelihood can be substantial even in the gaussian case for small samples if the autocorrelation of the gaussian process is high .
on the other hand , parametric methods are more powerful than nonparametric methods if the observed time series is close to the considered model class but fail if the model is misspecified . to exploit the advantages of both parametric and nonparametric approaches ,
the autoregressive - aided periodogram bootstrap has been developed by @xcite within the frequentist bootstrap world of time series analysis .
it fits a parametric working model to generate periodogram ordinates that mimic the essential features of the data and the weak dependence structure of the periodogram while a nonparametric correction is used to capture features not represented by the parametric fit .
this has been extended in various ways ( see @xcite ) .
its main underlying idea is a nonparametric correction of a parametric likelihood approximation .
the parametric model is used as a proxy for rough shape of the autocorrelation structure as well as the dependency structure between periodogram ordinates .
sensitivities with respect to the spectral density are mitigated through a nonparametric amendment .
we propose to use a similar nonparametrically corrected likelihood approximation as a pseudo - likelihood in the bayesian framework to compute the pseudo - posterior distribution of the power spectral density ( psd ) and other parameters in time series models .
this will yield a pseudo - likelihood that generalises the widely used whittle likelihood which , as we will show , can be regarded as a special case of a nonparametrically corrected likelihood under the gaussian i.i.d.working model .
software implementing the methodology is available in the ` r ` package ` beyondwhittle ` , which is available on the comprehensive r archive network ( cran ) , see @xcite .
the paper is structured as follows : in chapter [ sec : likelihood ] , we briefly revisit the whittle likelihood and demonstrate that it is a nonparametrically corrected likelihood , namely that of a gaussian i.i.d . working model .
then , we extend this nonparametric correction to a general parametric working model .
the corresponding pseudo - likelihood turns out to be equal to the true likelihood if the parametric working model is correctly specified but also still yields asymptotically unbiased periodogram ordinates if it is not correctly specified . in chapter
[ sec : bayesian ] , we propose a bayesian nonparametric approach to estimating the spectral density using the pseudo - posterior distribution induced by the corrected likelihood of a fixed parametric model .
we describe the gibbs sampling implementation for sampling from the pseudo - posterior .
this nonparametric approach is based on the bernstein polynomial prior of @xcite and used to estimate the spectral density via the whittle likelihood in @xcite .
we show posterior consistency of this approach and discuss how to incorporate the parametric working model in the bayesian inference procedure .
chapter [ sec : simulations ] gives results from a simulation study , including case studies of sunspot data , and gravitational wave data from the sixth science run of the laser interferometric gravitational wave observatory ( ligo ) .
this is followed by discussion in chapter [ sec : summary ] , which summarises the findings and points to directions for future work .
the proofs , the details about the bayesian autoregressive sampler as well as some additional simulation results are deferred to the appendices [ sec : proofs ] [ sec_appendixsims ] .
while the likelihood of a mean zero gaussian time series is completely characterised by its autocovariance function , its use for nonparametric frequentist inference is limited as it requires estimation in the space of positive definite covariance functions . similarly for nonparametric bayesian inference , it necessitates the specification of a prior on positive definite autocovariance functions which is a formidable task . a quick fix is to use parametric models such as arma models with data - dependent order selection , but these methods tend to produce biased results when the arma approximation to the underlying time series is poor .
a preferable nonparametric route is to exploit the correspondence of the autocovariance function and the spectral density via the wiener - khinchin theorem and nonparametrically estimate the spectral density . to this end , @xcite defined a pseudo - likelihood , known as the whittle likelihood , that directly depends on the spectral density rather than the autocovariance function and that gives a good approximation to the true gaussian and certain non - gaussian likelihoods . in the following subsection
we will revisit this approximate likelihood proposed by @xcite , before introducing a semiparametric approach which extends the whittle likelihood .
assume that @xmath0 is a real zero mean stationary time series with absolutely summable autocovariance function @xmath1 .
under these assumptions the spectral density of the time series exists and is given by the fourier transform ( ft ) of the autocovariance function @xmath2 consequently , there is a one - to - one - correspondence between the autocovariance function and the spectral density , and estimation of the spectral density is amenable to smoothing techniques .
the idea behind these smoothing techniques is the following observation , which also gives rise to the so - called whittle approximation of the likelihood of a time series : consider the periodogram of @xmath3 , @xmath4 the periodogram is given by the squared modulus of the discrete fourier coefficients , the fourier transformed time series evaluated at fourier frequencies @xmath5 , for @xmath6 .
it can be obtained by the following transformation : define for @xmath7 @xmath8 where @xmath9 and for @xmath10 even , @xmath11 is defined analogously . then , @xmath12 is an orthonormal @xmath13 matrix ( cf .
e.g.@xcite , paragraph 10.1 ) .
real- and imaginary parts of the discrete fourier coefficients are collected in the vector @xmath14 and the periodogram can be written as @xmath15 it is well known that the periodograms evaluated at two different fourier frequencies are asymptotically independent and have an asymptotic exponential distribution with mean equal to the spectral density , a statement that remains true for non - gaussian and even non - linear time series @xcite .
similarly , the fourier coefficients @xmath16 are asymptotically independent and normally distributed with variances equal to @xmath17 times the spectral density at the corresponding frequency .
this result gives rise to the following whittle approximation in the frequency domain @xmath18 by the likelihood of a gaussian vector with diagonal covariance matrix @xmath19 as explicitly shown in appendix [ sec : proofs ] , this yields the famous whittle likelihood in the time domain via the transformation theorem @xmath20 which provides an approximation of the true likelihood .
it is exact only for gaussian white noise in which case @xmath21 .
it has the advantage that it depends directly on the spectral density in contrast to the true likelihood that depends on @xmath22 indirectly via wiener - khinchin s theorem .
sometimes , the summands corresponding to @xmath23 as well as @xmath24 ( the latter for @xmath10 even ) are omitted in the likelihood approximation .
in fact , the term corresponding to @xmath23 contains the sample mean ( squared ) while the term corresponding to @xmath24 gives the alternating sample mean ( squared ) . both have somewhat different statistical properties and usually need to be considered separately .
furthermore , the first term is exactly zero if the methods are applied to time series that have been centered first , while the last one is approximately zero and asymptotically negligible ( refer also remark [ rem_firstlast ] ) . the density of @xmath25 under the i.i.d .
standard gaussian working model is the whittle likelihood .
it has two potential sources of approximation errors : the first one is the assumption of independence between fourier coefficients which holds only asymptotically but not exactly for a finite time series , the second one is the gaussianity assumption . in this paper , we restrict our attention to the first problem , extending the proposed methods to non - gaussian situations will be a focus of future work .
in fact , the independence assumption leads to asymptotically consistent results for gaussian data .
but even for gaussian data with relatively small sample sizes and relatively strong correlation the loss of efficiency of the nonparametric approach using the whittle likelihood can be substantial as shown in @xcite or by the simulation results of @xcite .
the central idea in this work is to extend the whittle likelihood by proceeding from a certain parametric working model ( with mean 0 ) for @xmath26 rather than an i.i.d .
standard gaussian working model before making a correction analogous to the whittle correction in the frequency domain .
to this end , we start with some parametric likelihood in the time domain , such as e.g. obtained from an arma - model , that is believed to be a reasonable approximation to the true time series .
we denote the spectral density that corresponds to this parametric working model by @xmath27 . if the model is misspecified , then this spectral density is also wrong and needs to be corrected to obtain the correct second - order dependence structure . to this end
, we define a correction matrix @xmath28 this is analogous to the whittle correction in the previous section as , in particular , @xmath29 with @xmath30 as in .
however , the corresponding periodogram ordinates are no longer independent under this likelihood but instead inherit the dependence structure from the original parametric model ( see proposition [ cor_31 ] c ) .
such an approach in a bootstrap context has been proposed and successfully applied by @xcite using an ar(@xmath31 ) approximation .
this concept of a nonparametric correction of a parametric time domain likelihood is illustrated in the schematic diagram : @xmath32 as a result we obtain the following nonparametrically corrected likelihood function under the parametric working model @xmath33 where @xmath34 denotes the parametric likelihood . [ rem_identifiability ] parametric models with a multiplicative scale parameter @xmath35 yield the same corrected likelihood as the one with @xmath36 , i.e. if @xmath37 is used as working model this leads to the same corrected likelihood for all @xmath38 .
for instance , if the parametric model is given by i.i.d .
@xmath39 random variables with @xmath40 arbitrary , then the correction also results in the whittle likelihood ( for a proof we refer to appendix [ sec : proofs ] ) .
analogously , for linear models @xmath41 , @xmath42 , which includes the class of arma - models , the corrected likelihood is independent of @xmath43
. we can now prove the following proposition which shows two important things : first , the corrected likelihood is the exact likelihood in case the parametric model is correct .
second , the periodograms associated with this likelihood are asymptotically unbiased for the true spectral density regardless of whether the parametric model is true . [ cor_31 ] let @xmath44 be a real zero mean stationary time series with absolutely summable autocovariance function @xmath45 and let @xmath46 for @xmath47 be the spectral density associated with the ( mean zero ) parametric model used for the correction . 1 . if @xmath48 , then @xmath49 .
the periodogram associated with the corrected likelihood is asymptotically unbiased for the true spectral density , i.e. @xmath50 where the convergence is uniform in @xmath51 .
furthermore , @xmath52 the proof shows that the vector of periodograms under the corrected likelihood has exactly the same distributional properties as the vector of the periodograms under the parametric likelihood multiplied with @xmath53 .
hence , asymptotic properties as the ones derived in theorem 10.3.2 in @xcite carry over with the appropriate multiplicative correction . in the remainder of the paper
we describe how to make use of this nonparametric correction in a bayesian set - up .
to illustrate the bayesian semiparametric approach and how to sample from the pseudo - posterior distribution , in the following we restrict our attention to an ar(@xmath31 ) model as our parametric working model for the time series , i.e. @xmath54 , where @xmath55 are i.i.d .
n(@xmath56 random variables with density denoted by @xmath57 .
note that without loss of generality , @xmath58 , cf .
remark [ rem_identifiability ] .
this yields the parametric likelihood of our working model , depending on the order @xmath59 and on the coefficients @xmath60 : @xmath61 with spectral density @xmath62 we assume the time series to be stationary and causal a priori .
thus , @xmath63 is restricted such that @xmath64 has no zeros inside the closed unit disc , c.f .
theorem 3.1.1 . in @xcite . for
now , we assume that the parameters @xmath65 of the parametric working model are fixed ( and in practice set to bayesian point estimates obtained from a preceding parametric estimation step ) .
an extension to combine the estimation of the parametric model with the nonparametric correction will be detailed later in section [ sec_priorparametric ] . for a bayesian analysis using
either the whittle or nonparametrically corrected likelihood , we need to specify a nonparametric prior distribution for the spectral density . here
we employ the approach by @xcite which is essentially based on the bernstein polynomial prior of @xcite as a nonparametric prior for a probability density on the unit interval .
we briefly describe the prior specification and refer to @xcite for further details .
in contrast to the approach in @xcite , we do not specify a nonparametric prior distribution for the spectral density @xmath66 , but for a pre - whitened version thereof , incorporating the spectral density of the parametric working model into the estimation .
to elaborate , for @xmath67 , consider the _ eta - damped correction function _
@xmath68 this corresponds to a reparametrization of the likelihood by replacing @xmath69 with @xmath70 .
[ rem_etacorrection ] the parameter @xmath71 models the confidence in the parametric model : if @xmath71 is close to @xmath72 and the model is well - specified , then @xmath73 will be much smoother than the original spectral density , since @xmath27 already captures the prominent spectral peaks of the data very well . as a consequence ,
nonparametric estimation of @xmath74 should involve less effort than nonparametric estimation of @xmath66 itself .
this remains true in the misspecified case , as long as the parametric model does describe the essential features of the data sufficiently well in the sense that it captures at least the more prominent peaks .
however , it is possible that the parametric model introduces erroneous spectral peaks if the model is misspecified . in that case
, @xmath71 close to zero ensures a damping of the model misspecification , such that nonparametric estimation of @xmath74 should involve less effort than nonparametric estimation of @xmath53 .
the choice of @xmath71 will be detailed in section [ sec_priorparametric ] , but for now , @xmath71 is assumed fixed .
we reparametrise @xmath74 to a density function @xmath75 on @xmath76 $ ] via @xmath77 with normalization constant @xmath78 .
thus , a prior for @xmath73 may be specified by putting a bernstein polynomial prior on @xmath75 and then an independent inverse - gamma@xmath79 prior on @xmath80 , its density denoted by @xmath81 .
the bernstein polynomial prior of @xmath82 is specified in a hierarchical way as follows : 1 .
@xmath83 \beta(\omega|j , k - j+1)$ ] where @xmath84=g(v)-g(u)$ ] for a distribution function @xmath85 and @xmath86 is the beta density with parameters @xmath87 and @xmath88 .
2 . @xmath85 has a dirichlet process distribution with base measure @xmath89 , where @xmath90 is a constant and @xmath91 a distribution function with lebesgue density @xmath92 .
3 . @xmath93 has a discrete distribution on the integers @xmath94 , independent of @xmath85 , with probability function @xmath95 .
note that smaller values of @xmath93 yield smoother densities .
furthermore , we achieve an approximate finite - dimensional characterization of this nonparametric prior in terms of @xmath96 parameters @xmath97 by employing the truncated sethuraman ( 1994 ) representation of the dirichlet process @xmath98 with @xmath99 , @xmath100 for @xmath101 , @xmath102 , and @xmath103 , all independent .
this gives a prior finite mixture representation of the eta - damped correction @xmath104 where @xmath105 and @xmath106 .
the joint prior density of @xmath107 by means of this finite - dimensional approximation can be written as @xmath108 here , we specify a diffuse prior by choosing the uniform distribution for @xmath91 and @xmath109 .
we set @xmath110 , @xmath111 and follow the recommendation by @xcite for the truncation point @xmath112 .
the prior on @xmath73 induces a prior on @xmath66 by multiplication with @xmath113 , see .
accordingly , the pseudo - posterior distribution of @xmath66 can be computed as prior times the corrected parametric likelihood : @xmath114 where @xmath115 and @xmath116 as in .
samples from the pseudo - posterior distribution can be obtained via gibbs sampling following the steps outlined in @xcite .
the full conditional for @xmath93 is discrete and readily sampled , as is the conjugate full conditional of @xmath80 .
we use the metropolis algorithm to sample from each of the full conditionals of @xmath117 and @xmath118 using the uniform proposal density of @xcite .
[ rem_firstlast ] as in @xcite , we omit the first and last terms in the corrected likelihood that correspond to @xmath119 and @xmath120 ( and setting @xmath121 as well as @xmath122 ) .
this is due to the role that the corresponding fourier coefficients play ( being equal to the sample mean respectively alternating sample mean ) , which typically requires a special treatment ( see proposition 10.3.1 and ( 10.4.7 ) in @xcite ) .
for the application to spectral density estimation in this paper this leads to more stable statistical procedures irrespective of the true mean of the time series .
however , in situations , where the time series is merely used as a nuisance parameter such as regression models , change point or unit - root testing , these coefficients should be included and the likelihood used for the time series @xmath123 , where @xmath124 is the mean ( not the sample mean ) of the time series . in this section
, we will show consistency of the pseudo - posterior distribution based on the bernstein polynomial prior and the corrected likelihood for a given working model under the same assumptions as @xcite . throughout the section
, we will make the following assumption : [ ass_1 ] 1 .
denote by @xmath125 and @xmath27 the autocovariance function respectively spectral density of the parametric working model .
assume that @xmath126 2 .
let @xmath44 be a stationary mean zero gaussian time series with autocovariance function @xmath127 and spectral density @xmath128 fulfilling @xmath129 denote by @xmath130 and @xmath131 the density and the distribution of @xmath132 .
an important first observation is , that the corrected likelihood , the whittle likelihood as well as the true likelihood are all mutually contiguous in the gaussian case .
this fact may also be of independent interest : [ theorem_contiguity ] under assumptions @xmath133.[ass_1 ] the true density @xmath130 , the whittle likelihood @xmath134 given in as well as the corrected ( gaussian ) parametric likelihood @xmath135 given in are all mutually contiguous . with the help of this theorem we are now able to prove posterior consistency under certain assumptions on the time series and prior .
[ theorem : consistency - corrected ] let @xmath136 fixed . let assumptions @xmath133.[ass_1 ] hold in addition to the following assumptions on the prior for @xmath73 : * for all @xmath93 , @xmath137 for some constants @xmath138 , * @xmath92 is bounded , continuous and bounded away from 0 , * the parameter @xmath80 is assumed fixed and known .
let @xmath139 .
then the posterior distribution is consistent , i.e. for any @xmath140 , @xmath141 in @xmath131-probability , where @xmath142 denotes the pseudoposterior distribution computed using the corrected likelihood . in the previous sections ,
the parameters of the working model were assumed to be fixed , as e.g. obtained in an initial pre - estimation step . from a bayesian perspective , it is desirable to couple the inference about the parametric working model with the nonparametric correction , allowing for the inclusion of prior knowledge about the model and for uncertainty quantification about the interaction of model and correction .
thus , for a fixed order @xmath31 , we include both the autoregressive parameters @xmath143 and the spectral shape confidence @xmath71 from into the bayesian inference .
the introduction of the parameter @xmath71 effectively robustifies the procedure in the sense that it guarantees our method will not be worse than a corresponding fully nonparametric one . to ensure stationarity and causality ( and hence identifiability ) of the parametric model
, we put a prior on the partial autocorrelations @xmath144 with @xmath145 for @xmath146 .
the autoregressive parameters @xmath147 can be readily obtained from this parametrisation ( see appendix [ sec_appendixsampling ] ) .
we consider the following prior specification for the spectral density : @xmath148 with a bernstein - dirichlet prior on @xmath73 as in section [ sec_bernsteindirichlet ] , a uniform prior on @xmath71 and uniform priors on the @xmath149 s , all a priori independent .
of course , it is possible to employ different prior models ( see @xcite ) . in conjunction with the corrected parametric likelihood
, we obtain samples from the joint pseudo - posterior distribution @xmath150 analogously to section [ sec_posteriorcomputation ] via gibbs sampling .
note that , since the corrected parametric likelihood is the lebesgue density of a probability measure , it is sufficient that the prior distributions are proper for the posterior distribution to be proper .
we use random walk metropolis - within - gibbs steps with normal proposal densities to sample from the full conditionals of @xmath71 and @xmath151 respectively . the proposal variance for @xmath71
is set to 0.01 , where proposals larger than 1 ( smaller than 0 ) are truncated at 1 ( at 0 ) . to achieve proper mixing of the parametric model parameters , the proposal variances @xmath152 for @xmath149
are determined adaptively as described in @xcite during the burn - in period , aiming for an acceptance rate of 0.44 , where proposals with absolute value larger or equal to one are discarded .
[ rem_arorder ] the autoregressive order @xmath31 is assumed to be fixed . in our approach , it is determined in a preliminary model selection step
. however , it is also possible to include the autoregressive order in the bayesian inference by using a reversible - jump markov chain monte carlo scheme @xcite or stochastic search variables @xcite .
in this section , we evaluate the finite sample behavior of our _ nonparametrically corrected ( npc ) _ approach to bayesian spectral density estimation numerically .
to demonstrate the trade - off between the parametric working model and the nonparametric spectral correction , we compare our approach to both fully parametric and fully nonparametric approaches .
we first present the results of a simulation study with arma data in section [ sec : simarma ] before considering sunspot data in section [ sec : sunspot ] and gravitational wave data in section [ sec : ligo ] .
an implementation of all procedures presented below is provided in the ` r ` package ` beyondwhittle ` , which is available on cran , see @xcite .
we consider data generated from the arma model @xmath153 with standard gaussian white noise @xmath154 and different values of @xmath155 and @xmath10 .
the following competing approaches are compared with npc : _ nonparametric estimation ( np ) . _
the procedure from @xcite , which is based on the whittle likelihood and a bernstein - dirichlet prior on the spectral density . note that this coincides with the npc approach with a white noise parametric working model ( @xmath156 ) , c.f .
remark [ rem_identifiability ] .
_ autoregressive estimation ( ar ) .
_ for @xmath157 , an autoregressive model of order @xmath31 is fitted to the data using a bayesian approach with the same partial autocorrelation parametrization and the same prior assumptions as for the parametric working model within the nonparametrically corrected likelihood procedure , see section [ sec_priorparametric ] ( for details on the sampling scheme we refer to appendix [ sec_appendixsampling ] ) .
the order @xmath158 minimizing the dic is then chosen for model comparison . the working model in the npc approach
is chosen to be the ar(@xmath158 ) model from the ar procedure .
the prior for the working model parameters is as described in section [ sec_priorparametric ] and the prior for the nonparametric correction is as described in section [ sec_bernsteindirichlet ] . for the np approach ,
the same bernstein - dirichlet prior for @xmath66 is used as for @xmath73 in the npc approach .
we compare the average integrated absolute error ( aiae ) of the posterior median spectral density estimate and the empirical coverage probability of a uniform credible interval ( cuci ) .
note that pointwise posterior credible intervals are not suited for investigating coverage , since they do not take the multiple testing problem into account that arises at different frequencies .
following @xcite ( see also @xcite ) , a uniform credible interval for the spectral density can be constructed as follows : denote by @xmath159 the posterior spectral density samples obtained by one of the procedures . then for @xmath160 the uniform @xmath161-credible interval is given by @xmath162\ ] ] where @xmath163 denotes the sample median at frequency @xmath164 , @xmath165
the median absolute deviation of @xmath166 and @xmath167 is chosen such that @xmath168 the intervals are constructed on a logarithmic scale to ensure that their covered range contains only positive values .
because small values of @xmath169 lead to very large absolute values on a log scale , we do not employ the usual logarithm , but the fuller - logarithm as described in @xcite , page 496 , i.e. @xmath170 for some small @xmath171 .
we use @xmath172 and @xmath173 in our simulations .
the chains were run for 12,000 iterations for ar ( after a burn - in period of 8,000 iterations ) and for 20,000 iterations for np and npc ( after a burn - in period of 30,000 iterations ) , where a thinning of 4 was employed for np and npc .
we choose @xmath174 and consider lengths @xmath175 from model with @xmath176 replicates ( @xmath177 a power of 2 to use the computational resources efficiently ) respectively .
the results are shown in table 1 . for ar(1 ) data , the ar procedure yields the best results ( in terms of both aiae and cuci ) , whereas the np performs worst .
it can be seen that ar and npc benefit from the well - specified parametric model . for ma(1 )
data , however , the ar approach yields the worst results , whereas npc benefits from the nonparametric correction , yielding only slighly worse integrated errors than np , although with superior uniform credible intervals for @xmath178 . in case of arma(1,1 ) data , the estimation does not benefit from the autoregressive fit , i.e. the moving average misspecification dominates the estimation .
thus the results are similar to the ma(1 ) case .
further results for data from the arma model can be found in appendix [ sec_appendixsims ] .
max width= [ rem_visualinspection ] under relatively weak conditions ( see e.g. @xcite ) a linear process can be written as an ar(@xmath179)-process with white noise errors ( similarly to the famous wold representation ) .
consequently , an ar(@xmath31)-model with sufficiently large order captures the structure of a ( gaussian ) linear process to any degree of accuracy . in this sense , the use of an ar - model with a sufficiently large order can be viewed as a nonparametric procedure , a fact , that has been exploited by ar - sieve - bootstrap methods for quite some time .
for a recent mathematical analysis of the validity and limitations of this approach we refer to @xcite .
consequently , an ar - model can still be used for spectral density estimation under misspecification as long as the order is sufficiently large . in this sense
standard order selection techniques such as dic - minimization tend to choose large orders in this situation . however , looking at scree - like plots of the negative maximum log likelihood for increasing orders one can often see a clear bend ( elbow ) in the curve ( with a slow decay from that point on that is not slow enough to be captured by standard penalization techniques ) .
similar to the use of scree plots in the context of pca , that point can be seen as a reasonable truncation point ( elbow criterion ) where those features best explained by the parametric model have been captured . while this small model does not yet fully explain the data , adding more parameters is not helping the nonparametrically corrected procedure that we propose .
in other words , we are not interested in an elaborate ar(@xmath179 ) approximation but rather in a proxy model that captures the main parametric features of the data . in the context of an autoregressive working model , we approximate the negative maximum log - likelihood by the negative log - likelihood evaluated at the yule - walker estimate .
this is to ensure numerical stability and computational speed , especially for large orders .
the approximation is motivated by the asymptotic equivalence of both estimates , see e.g. chapter 8 in @xcite .
the estimate is referred to as _ negative maximum log - likelihood _ in the text .
figure [ fig : visualinspectionarma ] shows the scree - like plots for three exemplary arma(1,1 ) realizations from the above model as well as the sunspot data set . in all three realizations the elbow is clearly at @xmath180 ( which is consistent with the ar - part of the model ) , while the dic - criterion choses orders between 4 and 6 .
table 2 shows the simulation results for the arma(1,1 ) model and a fixed order of @xmath180 . while a parametric ar(@xmath72 ) model is clearly not able to explain the data ( see e.g. the zero coverage of the uniform credibility intervals ) , this choice of the order significantly improves the results of the npc procedure for the arma(1,1 ) data . in fact , the latter is now better than both the ar procedure as well as the np procedure while at the same time the confidence in the model as indicated by @xmath181 increases .
max width= for the sunspot data that effect can also be seen clearly as the above procedure proposes to use @xmath182 in the nonparametric procedure while the dic - criterion suggests @xmath183 . for a detailed discussion of this data analysis
we refer to section [ sec : sunspot ] , similar effects for the ligo data are discussed in section [ sec : ligo ] . in this section , we analyse the yearly sunspot data from 1700 until 1987 .
we take the mean - centered version of the square root of the 288 observations as input data .
we compare the ar and the npc procedure for fixed values @xmath184 .
while @xmath183 minimises the dic , @xmath182 captures the main ar - features of the data as indicated by the elbow criterion ( see remark [ rem_visualinspection ] and figure [ fig : visualinspectionarma ] ( d ) ) .
the results are shown in figure [ fig_sunspot ] . .
the log - periodogram is visualised in grey . ]
while for @xmath180 the bernstein polynomials of the nonparametric correction can not yet capture the peaks sufficiently well , this is clearly the case for @xmath182 ( the choice obtained from the elbow criterion ) : while the parametric model itself can clearly not yet explain the data well , enough features are captured to improve the nonparametric correction . for larger order choices , the estimate from the npc method does not change much anymore , so that the correction does indeed not profit from additional parameters in the ar - model .
in fact for @xmath183 ( as indicated by dic ) , the bayes estimator of ar and npc are very similar . gravitational waves , ripples in the fabric of spacetime caused by accelerating massive objects , were predicted by albert einstein in 1916 as a consequence of his general theory of relativity , see @xcite .
gravitational waves originate from non - spherical acceleration of mass - energy distributions , such as binary inspiraling black holes , pulsars , and core collapse supernovae , propagating outwards from the source at the speed of light .
however , they are very small ( a thousand times smaller than the diameter of a proton ) so that their measurement has provided decades of enormous engineering challenges . on sept .
14 , 2015 , the laser interferometric gravitational wave observatory ( ligo ) , see @xcite , made the first direct detection of a gravitational wave signal , gw150914 , originating from a binary black hole merger @xcite .
the two l - shaped ligo instruments ( in hanford , washington and livingston , louisiana ) each consist of two perpendicular arms , each 4 kilometers long .
a passing gravitational wave will alternately stretch one arm and squeeze the other , generating an interference pattern which is measured by photo - detectors .
the detector output is a time series that consists of the time - varying dimensionless strain @xmath185 , the relative change in spacing between two test masses .
the strain can be modelled as a deterministic gravitational wave signal @xmath186 depending on a vector @xmath187 of unknown waveform parameters plus additive noise @xmath188 , such that @xmath189 there are a variety of noise sources at the ligo detectors .
this includes _
seismic _ noise , due to the motion of the mirrors from ground vibrations , earthquakes , wind , ocean waves , and vehicle traffic , _
thermal _ noise , from the microscopic fluctuations of the individual atoms in the mirrors and their suspensions , and _ shot _ noise , due to the discrete nature of light and the statistical uncertainty from the `` photon counting '' that is performed by the photo - detectors .
in particular , ligo noise includes high power , narrow band , spectral lines , visible as sharp peaks in the log - periodogram .
as the ligo spectrum is time - varying and subject to short - duration large - amplitude transient noise events , so - called `` glitches '' , a precise and realistic modelling and estimation of the noise component jointly with the signal is important for an accurate inference of the signal parameters @xmath187 .
the current approach , which was also used for estimating the parameters of gw150914 in @xcite , is to first use the welch method @xcite to estimate the spectral density from a separate stretch of data , close to but not including the signal and then to assume stationary gaussian noise with this known spectral density in order to estimate the signal parameters .
several approaches have been suggested in the recent gravitational wave literature to simultaneously estimate the noise spectral density and signal parameters .
these include generalising the whittle likelihood to a student - t likelihood as in @xcite , similarly modifying the likelihood to include additional scale parameters and then marginalising over the uncertainty in the psd as in @xcite , using cubic splines for smoothly varying broad - band noise and lorentzians for narrow - band line features as in @xcite , a morlet - gabor continuous wavelet basis for both gravitational wave burst signals and glitches as in @xcite , the nonparametric approach of @xcite using a dirichlet - bernstein prior @xcite and a generalisation of this using a b - spline prior , see @xcite .
we consider 1 s of real ligo data collected during the sixth science run ( s6 ) , recoloured to match the target noise sensitivity of advanced ligo @xcite .
the data is differenced and then multiplied by a hann window to mitigate spectral leakage .
a low - pass butterworth filter ( of order 20 and attenuation 0.25 ) is then applied before downsampling from a ligo sampling rate of 16384 hz to 4096 hz , reducing the volume of data . )
models applied to advanced ligo s6 data . ]
we first run a pure nonparametric model , corresponding to a nonparametrically corrected likelihood with an ar(0 ) working model ( i.e. the whittle likelihood ) to estimate the spectral density .
we then compare this to a nonparametrically corrected model with an order of @xmath190 , where a clear elbow can be seen in the negative log - likelihood plot ( see figure [ fig_screeligo ] and remark [ rem_visualinspection ] ) .
we run these simulations for 100,000 mcmc iterations , with a burn - in of 50,000 , and thinning factor of 5 .
results are illustrated in figure [ fig_gw_npc ] ( a ) .
even though @xmath93 converged to @xmath191 mixture components , it is clear that the bernstein - dirichlet prior together with the whittle likelihood is not flexible enough to estimate the sharp peaks of the ligo spectral density .
the parametric ar(14 ) model ( estimated using the bayesian autoregressive sampler described in appendix [ sec_appendixsampling ] ) captures the four main peaks but not their sharpness .
additionally , it does not capture the structure well in the frequency bands 0 to 450 hz as well as larger than 1100 hz . when compared to the ar(0 ) model , the nonparametrically corrected model based on @xmath192 estimates the sharp peaks much better .
furthermore , it sharpens all four peaks of the ar(14)-model ( with a slight exception around 400 hz , where seemingly two very sharp peaks overlap , a feature that is not captured by the ar(14 ) model at all ) . in the frequency bands 0 to 450 hz as well as larger than 1100 hz , where the parametric model fails altogether , the correction yields similar results to the nonparametric whittle procedure . similarly to the nonparametric whittle procedure @xmath93 tends towards @xmath193 indicating that the bernstein - dirichlet prior together with an ar(@xmath194)-model is not yet flexible enough for this data set .
looking closer at figure [ fig_screeligo ] , the negative log - likelihood between the models ar(14 ) and ar(35 ) decreases by 533 , which is not as sharp as the elbow at @xmath192 , but still significant keeping in mind that the ligo data is a very complex data set
much more so than the arma(1,1 ) or the sunspot data .
there is a moderately sized jump before @xmath195 , while afterwards the descend slows down significantly .
in fact the bic chooses an order of @xmath196 , where the log - likelihood reaches the level of 879 , showing that the difference between @xmath192 and @xmath195 ( of 533 ) is comparable to the one between @xmath195 and @xmath197 ( of 512 ) .
this indicates that there is indeed another change of gradient around @xmath195 .
when looking at penalized likelihoods , this is also the point , where different penalizations start to obviously diverge .
the results for @xmath195 can be found in figure [ fig_gw_npc ] ( b ) .
the parametric ar(@xmath198 ) model already provides a reasonable fit to the periodogram , picking up the major peaks ( with the exception around 100 hz ) , but under- and overestimates some of the peaks .
in particular there are still major problems in the frequency bands 0 to 300 hz and 400 to 700 hz .
the npc procedure with @xmath195 keeps the peaks that have been capture well by the parametric model but corrects problems most prominently in the above mentioned frequency bands .
it is worth mentioning that the correction works in several ways : sharpening existing peaks ( e.g. at 0 hz ) , adding new peaks ( e.g. at 100 hz ) as well as smoothing out some erroneous peaks ( e.g. at 600 hz ) .
overall , the resulting estimate seems to capture the structure quite well in all frequency bands . this impression is complemented by the results of the npc method with ar(@xmath198 ) working model together with the pointwise and uniform credible bounds obtained from the procedure in figure [ fig_gw_ci ] .
in this work we propose a nonparametric correction of a parametric likelihood to obtain an approximation of the true ( inaccessible ) likelihood of a stationary time series .
this approach extends the famous approximation by whittle . for gaussian data , the whittle likelihood , the nonparametric correction as well as the true likelihood are asymptotically equivalent .
secondly , we propose a bayesian procedure for spectral density estimation , where a parametric likelihood is used with a bayesian nonparametric spectral correction .
we show consistency of the resulting pseudo - posterior distribution for a fixed parametric likelihood .
furthermore , we present a bayesian semiparametric procedure that combines inference about the parametric working model with the nonparametric correction .
the extent of the contribution of the parametric spectral density to the spectral density estimate is controlled by a shape confidence parameter .
simulation results have shown that this procedure inherits the benefits from the parametric working model if the latter is well - specified or describe a part of the features of the data well , while in the misspecified case the results are comparable to the usage of the whittle likelihood . regarding future work , it is interesting to investigate whether in the non - gaussian case the class of time series for which asymptotically consistent inference holds can be enlarged by choosing an appropriate model .
it is important to understand the distributional influence in the non - gaussian case both in finite samples and asymptotically . as an example , in a bootstrap context , it suffices to capture the fourth order structure of a linear model to obtain asymptotically valid second - order frequentist properties of the autocovariance structure @xcite . as suggested by @xcite in a parametric
setting , this property does not simply carry over to a bayesian context .
preliminary results for non - gaussian autoregressive time series however have shown considerable benefits ( with respect to first and second order frequentist properties ) when the innovation distribution is well - specified in comparison to a gaussian model .
since any parametric likelihood is susceptible to misspecification , the ultimate goal is to consider a bayesian nonparametric model for the innovation distribution , such as dirichlet mixtures of normals .
further directions for future work are automation of the elbow criterion discussed in remark [ rem_visualinspection ] . instead of choosing a fixed order in advance , an automation might as well serve as a guideline for specifying a prior on the ar order parameter , which can be included in the inference by means of rjmcmc ( c.f .
remark [ rem_arorder ] ) .
this work was supported by dfg grant ki 1443/3 - 1 .
furthermore , the research was initiated during a visit of the fourth author at karlsruhe institute of technology ( kit ) , which was financed by the german academic exchange service ( daad ) .
some of the preliminary research was conducted while the first author was at kit , where her position was financed by the stifterverband fr die deutsche wissenschaft by funds of the claussen - simon - trust .
we also thank the new zealand escience infrastructure ( nesi ) and the universittsrechenzentrum ( urz ) magdeburg for their high performance computing facilities , and the centre for eresearch at the university of auckland and jrg schulenburg for their technical support .
since @xmath199 is orthonormal it holds @xmath200 and hence by an application of the transformation theorem @xmath201 where @xmath202 . finally , by , @xmath203 as well as @xmath204 it holds @xmath205 yielding the assertion .
the spectral density corresponding to a gaussian white noise with variance @xmath43 is given by @xmath206 , hence @xmath207 and @xmath208 which can be shown to be the whittle likelihood analogously to the proof of equation [ eq_whittle ] since @xmath209 . if @xmath48 then @xmath210 and @xmath211 , hence a ) follows .
for b ) , consider a time series distributed according to the corrected likelihood , i.e. @xmath212 .
an application of the transformation theorem shows that @xmath213 on noting that @xmath214 and @xmath215 . by
it holds @xmath216 as @xmath217 . by @xcite , proposition 10.3.1 . , it holds @xmath218 where convergence is uniform in @xmath219 ( recall that the time series is mean zero ) . from this
the assertion follows because @xmath220 is bounded from below by assumption .
@xcite proved mutual contiguity of the true gaussian and the whittle likelihood in the frequency domain which carries over to the time domain by an application of the transformation theorem because @xmath199 is bijective .
hence , it is sufficient to show mutual contiguity of the corrected parametric likelihood and the whittle likelihood . following the proof of theorem 1 in @xcite it is enough to show that their log - likelihood ratio is a tight sequence under both @xmath221 as well as @xmath222 . to this end , note that it holds @xmath223 where @xmath224 is the covariance matrix of the corresponding parametric time series , e.g. for the ar(@xmath31)-case the covariance matrix associated with likelihood .
hence , the log - likelihood ratio is given by @xmath225 defining @xmath226 analogously to @xmath30 as in with the nonparametric spectral density @xmath22 replaced by the parametric version @xmath220 as e.g. for ar(@xmath31 ) given in , we get @xmath227 where the boundedness follows from lemma a.1 in @xcite .
to obtain stochastic boundedness of @xmath228 under @xmath222 as well as @xmath221 we will show boundedness of the expectation and variance .
following @xcite we get under @xmath222 ( i.e. @xmath229 ) @xmath230 because @xmath231 , it holds by @xmath232 @xmath233 the assertion follows by lemma a.2 in @xcite by the linearity of the trace .
similar arguments yield the assertion under @xmath221 ( i.e. @xmath234 noting that @xmath235 concerning the variance we get under @xmath222 @xmath236 similar arguments as above yield @xmath237 together with this yields @xmath238 by @xmath239 .
hence , the assertion follows by lemma a.2 in @xcite .
analogous assertions yield the result under @xmath221 .
@xcite give sufficient conditions for the consistency of the posterior distribution when using bernstein polynomial priors in terms of the existence of exponentially powerful tests for testing @xmath240 and prior positivity of a kullback - leibler neighbourhood .
but these require i.i.d . observations . to prove posterior consistency under the whittle likelihood ,
@xcite extend this result to independent but not identically distributed observations and apply this to the periodogram ordinates which are independent exponential random variables under the whittle likelihood . however , periodogram ordinates under the corrected likelihood are no longer independent , therefore this theorem is not applicable .
we give an extension to non - independent random variables in the following theorem , which is needed to prove theorem [ theorem : consistency - corrected ] : [ theorem : consistency - general ] let @xmath241 be random vectors with probability distribution @xmath242 and corresponding pdf @xmath243 .
let @xmath244 , @xmath245 , where @xmath246 denotes the borel @xmath247-algebra on @xmath248 , and @xmath249 a probability distribution on @xmath250 .
define @xmath251 \mbox { and } \\
v_n(\theta_0,\theta ) & = & { \operatorname{var}}_{\theta_0}\left [ \log \frac{p_n({{\mathbf{z}}}_n|\theta_0)}{p_n({{\mathbf{z}}}_n|\theta ) } \right ] .\end{aligned}\ ] ] under the following assumptions of prior positivity of neighbourhoods and existence of tests : * there exists a set @xmath252 with @xmath253 such that * * @xmath254 , and * * @xmath255 .
* there exists test functions @xmath256 , subsets @xmath257 , and constants @xmath258 such that * * @xmath259 .
* * @xmath260 , and * * @xmath261 then @xmath262 the proof is completely analogous to the proof of theorem a.1 in @xcite
. the following lemma replaces lemma b.3 in @xcite .
[ lemma_b3 ] let @xmath263 where @xmath264 is a symmetric positive definite matrix and @xmath265 .
with @xmath266 consider testing @xmath267 where @xmath268 do not depend on @xmath10
. then there exists a test @xmath269 and constants @xmath270 depending only on @xmath271 and @xmath272 such that @xmath273 consider a test @xmath269 that rejects @xmath274 if @xmath275 with a critical value @xmath276 with @xmath277 and @xmath278 . denote @xmath279 , then @xmath280 consequently , the moment generating function of @xmath281 under @xmath274 is given by @xmath282=(1 - 2t)^{-n/2}$ ] and exists for @xmath283 . by an application of the markov inequality , we get for all @xmath284 @xmath285=e^{-n [ -zx+\frac 1 2 \log(1 + 2z)]}.\end{aligned}\ ] ] the function @xmath286 attains its maximum at @xmath287 and @xmath288 as @xmath289 for @xmath290 . thus ,
setting @xmath291 , we obtain @xmath292 similarly , under @xmath293 , we get @xmath294 since @xmath295s_n^{-1/2})=\mbox{det}(d_n^{1/2}(f_1/f_0)-\lambda i_n)\end{aligned}\ ] ] the matrix @xmath296 has the eigenvalues @xmath297 , @xmath298 .
since @xmath296 is a normal matrix ( recall that @xmath264 is symmetric positive definite ) , we find that @xmath299 has the eigenvalues @xmath300 , @xmath298 .
consequently , @xmath301 consequently , @xmath302 analogously to the proof under the null hypothesis , we get for any @xmath303 @xmath304 now @xmath305 , @xmath306 , attains its maximum at @xmath307 with @xmath308 as @xmath309 for @xmath310 , i.e. @xmath278 . setting @xmath311 yields @xmath312 , completing the proof .
we follow @xcite , proof of theorem 1 , and show ( c1 ) and ( c2 ) of theorem [ theorem : consistency - general ] above .
let @xmath313 denote the corrected likelihood and define @xmath314}|c_{0,\eta}(\lambda)|$ ] and @xmath315 } |c_{0,\eta}(\lambda)|$ ] .
an analogous argument as in appendix b.1 .
of @xcite shows that for all @xmath140 the set @xmath316 has positive prior probability under the bernstein polynomial prior on @xmath248 .
we need to show ( c1)(a ) and ( b ) . to this end , let @xmath317 where @xmath318 for @xmath317 we have @xmath319 . to prove ( c1 ) ,
first note that @xmath320 where @xmath321 .
for @xmath322 it holds @xmath323 as well as @xmath324 .
furthermore , by analogous argument as in the proof of lemma [ lemma_b3 ] we get that the eigenvalues of @xmath325 are given by @xmath326 , @xmath298 , hence @xmath327 as a consequence , we get @xmath328 similar arguments yield @xmath329\\ & = \frac{1}{2n^2}\sum_{i=1}^n\left ( \frac{c_{0,\eta}(\lambda_i)-c(\lambda_i)}{c(\lambda_i ) } \right)^2 = o(1/n)\;\|c - c_{0,\eta}\|_{\infty}^2 \to 0.\end{aligned}\ ] ] the proof can now be concluded as in @xcite by replacing lemma b.3 by lemma [ lemma_b3 ] above .
for fixed order @xmath330 , the autoregressive model @xmath331 with i.i.d .
n(@xmath332 ) random variables is parametrized by the the innovation variance @xmath43 and the partial autocorrelation structure @xmath333 , where @xmath334 is the conditional correlation between @xmath335 and @xmath336 given @xmath337 .
note that @xmath338 and that there is a one - to - one relation between @xmath339 and @xmath340 .
to elaborate , we follow @xcite and introduce the auxiliary variables @xmath341 as solutions of the following yule - walker - type equation with the autocorrelations @xmath342/{\operatorname{e}}z_1 ^ 2 $ ] : @xmath343 as shown in @xcite , the well - known relationships @xmath344 readily imply @xmath345 and ( recall that @xmath346 for @xmath347 ) the recursive formula @xmath348 we specify the following prior assumptions for the model parameters : @xmath349 all a priori independent .
we use @xmath350 .
furthermore , we employ the gaussian likelihood @xmath351 with the n(@xmath332 ) density @xmath352 and the @xmath353 autocovariance matrix @xmath354 of the autoregressive model . to draw samples from the corresponding posterior distribution
, we use a gibbs sampler , where the conjugate full conditional for @xmath43 is readily sampled . for
@xmath355 the full conditional for @xmath356 is sampled with the metropolis algorithm using a normal random walk proposal density with proposal variance @xmath152 . as for the parameters of the working model within the corrected parametric approach ( see section [ sec_priorparametric ] ) , the proposal variances are adjusted adaptively during the burn - in period , aiming for a respective acceptance rate of 0.44 .
table 2 depicts further results of the ar , np and npc procedures for normal arma data . for white noise ,
all procedures yield good results , whereas np is superior due to the implicitly well - specified white noise working model ( recall that the order @xmath31 within ar and npc are estimated with the dic ) . the results for ma(2 )
data are qualitatively similar to the results for ma(1 ) data in table 1 .
the spectral peaks of the ar(2 ) model are not as strong as the spectral peak of the ar(1 ) model considered in table 1 .
accordingly , it can be seen that the np results are better in this case .
the npc benefits again from the well - specified parametric model , yielding results that are comparable to the ar procedure . | * abstract . * the whittle likelihood is widely used for bayesian nonparametric estimation of the spectral density of stationary time series .
however , the loss of efficiency for non - gaussian time series can be substantial .
on the other hand , parametric methods are more powerful if the model is well - specified , but may fail entirely otherwise .
therefore , we suggest a nonparametric correction of a parametric likelihood taking advantage of the efficiency of parametric models while mitigating sensitivities through a nonparametric amendment . using a bernstein - dirichlet prior for the nonparametric spectral correction , we show posterior consistency and illustrate the performance of our procedure in a simulation study and with ligo gravitational wave data . |
observations of protoplanetary disks the birthplace of planets provide a window into the processes by which planets form . while the dust component of protoplanetary disks has historically been easier to observe , the gas component contains the bulk of the disk mass , and encodes much of the important disk physics and chemistry . therefore , obtaining observations of the disk gas remains an important goal , in spite of its challenges . to date , the majority of molecular gas detections in disks has come from infrared ( ir ) observations , where the heated disk atmospheres in the inner parts of disks ( typically within @xmath55 au ) produce emission from rovibrational or high - energy pure rotational molecular transitions ( e.g. * ? ? ?
* ; * ? ? ?
considerably fewer observations have been obtained of disk gas from the outer parts of disks , especially observations of the lowest - energy pure rotational transitions at millimeter wavelengths , as the sensitivity of current instruments allowed for the high - resolution study of only the brightest protoplanetary disks ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
yet it is the combination of observations across a wide range of wavelengths , probing a wide range of disk radii , that will enable a more complete understanding of disks .
this situation is about to change dramatically , as the superior sensitivity and image quality provided by the atacama large millimeter array ( alma ) is enabling extensive observations of disk gas at millimeter wavelengths . in particular
, alma observations will allow for detailed studies of its motion and distribution .
indeed , alma observations have already revealed gaseous flows across a disk cavity @xcite , spirals @xcite and a warped velocity field , consistent with fast radial inflow @xcite , in the transitional disk hd 142527 , as well as an inner clearing in the transitional disk ophiuchus irs 48 @xcite .
alma will also allow for the study of winds and outflows from disks , probing more tenuous and smaller - scale regions than ever before ; a helical disk wind has already been detected around the hd 163296 gas disk @xcite .
the work presented here was inspired by an observational puzzle derived from ir observations .
for all moderate or high inclination disks , the keplerian motion of the gas molecules is expected to produce double - peaked emission line profiles for the rovibrational lines produced within a few au of the central star . surprisingly , however , observations have revealed that co rovibrational emission lines from some moderate - inclination disks are strongly single - peaked ( as defined by the so - called line profile parameter ; @xcite ) .
as they also have a small spatial extent , these emission lines are inconsistent with any keplerian disk models .
furthermore , spectro - astrometry reveals that the lines can not be explained solely as emission from an azimuthally - symmetric keplerian disk @xcite . as a group
, the disks with strongly single - peaked emission lines also tend to have high rovibrational co line / continuum ratios , and signatures of high accretion rates ( @xcite ) , including high uv fluxes @xcite and high optical and ir veiling ( e.g. * ? ? ?
* ; * ? ? ?
in addition , several disks with single - peaked emission lines are known to have jets or outflows @xcite .
strongly single - peaked profiles are seen in perhaps @xmath530% of co rovibrational line - emitting disks ( @xcite ) , but recent results suggest that nearly all protoplanetary disks produce emission lines that are more single - peaked than expected for a keplerian disk , and that they tend to have a slight blue asymmetry @xcite . therefore , whatever phenomenon explains these observations may be widespread .
a possible explanation for these surprising observations was suggested by @xcite , who showed that disk + disk wind models could reproduce the observed lineshapes , spectro - astrometric line profiles and small spatial extent .
while the models used by @xcite were originally derived from magneto - hydrodynamic wind simulations @xcite , they were largely parameterized , and only loosely constrained by the available data . yet
, if the disk wind models used by @xcite are correct , disks may be losing mass at a rate of @xmath5 1% of their mass accretion rate , and the results of @xcite suggest that disk winds may in fact be universal .
how might these winds affect disk evolution , or the formation of planets ? the answers to these questions are essentially unknown , and require that we gain a better understanding of the properties and driving mechanisms of these winds .
if the disk wind cools quickly as it extends outwards , and its density drops , a low - critical - density , low - excitation tracer like @xmath1co @xmath6 may be a good choice for detecting a wind on larger scales , and for studying its structure and kinematics . in this work ,
we present alma observations of co @xmath6 from a prototypical source displaying evidence for a possible disk wind .
the target of our study the t tauri star as 205 n shows strongly centrally peaked ir molecular emission lines with high line / continuum ratios @xcite , a high and possibly highly variable accretion rate @xcite , and some evidence for an outflow @xcite . as
205 n is also the northern and more massive component of a @xmath7 ( projected separation ) binary system , in which the secondary star ( as 205 s ) is itself a double - lined spectroscopic binary @xcite . as 205
is located at the northern edge of the ophiuchus star forming region , at a distance of @xmath5125 pc @xcite . in section
[ sec : observations ] , we briefly describe the data acquisition and reduction . in section [ sec : results ] , we present detections of @xmath1co , @xmath2co and c@xmath3o from the disks in the as 205 binary system , and discuss interesting features in the images and velocity fields .
then , in section [ sec : origin ] we discuss possible explanations for these features , including an extensive discussion of their possible connection to the disk winds proposed to explain as 205 n s unusual ir spectroscopic features .
finally , we include a brief disussion of implications ( section [ sec : discussion ] ) and conclusions ( section [ sec : conclusions ] ) .
observations of as 205 n and s were obtained on ut 27 mar 2012 and 04 may 2012 , with the cycle 0 extended configuration .
the nearby quasar j1625 - 254 , at a distance of @xmath8 , was observed at @xmath511 minute intervals to correct for atmospheric variations .
short - term phase variations were corrected using the on - board water vapor radiometers .
fifteen 12-meter antennas were available for both sets of observations , although one to two antennas were flagged at any given time during the observations . on 27 mar , the minimum and maximum baselines were 43 and 402 meters ( corresponding to minimum and maximum spatial scales of @xmath9 and @xmath10 , 88 and 500 au at 125 pc ) , respectively , and on 04 may , they were 21 and 402 meters ( with a maximum spatial scale of @xmath11 au ) .
the total on - source integration time for each set of observations was 15.8 minutes , for a total integration time of 31.6 minutes .
observations were obtained in four spectral windows of width 117 mhz , with a channel width of 30.5 khz ( 0.04 km s@xmath12 ) , corresponding to a spectral resolution of 61 khz ( 0.08 km s@xmath12 ) .
three of the spectral windows were centered at the frequencies of @xmath2co @xmath13 ( 220.399 ghz ) , @xmath1co @xmath13 ( 230.538 ghz ) and c@xmath3o @xmath13 ( 219.56 ghz ) .
the fourth window was centered on an undetected ch@xmath14oh transition ( 232.415 ghz ) , and was used for continuum measurement .
the data were reduced according to a standard procedure , using the common astronomy software applications package ( casa ; @xcite ) .
reduction steps included passband calibration , correction for atmospheric variations , and absolute flux calibration .
absolute flux was calibrated using observations and models of titan .
the calibrated measurement sets were then combined for the two dates .
the continuum window was binned to a single channel , then cleaned using the clark algorithm and briggs weighting with a robustness parameter of 0.5 .
the spectral windows were first continuum - subtracted , then binned to a channel spacing equal to the spectral resolution ( 0.08 km s@xmath12 ) and cleaned in the same manner as the continuum window . continuum subtraction was performed with the casa routine _ uvcontsub _ using a fit to the 232 and 219 ghz windows and either a constant model ( for @xmath1co ) or a linear model ( for @xmath2co and c@xmath3o ) .
inspection of line - free channels suggests continuum - subtraction residuals in the @xmath1co maps of up to @xmath510% near the centers of as 205 n and s ; no significant residuals remain in the @xmath2co and c@xmath3o maps .
in figure [ fig : map_plot ] , we show the observed ( cleaned ) @xmath1co @xmath13 , @xmath2co @xmath13 and c@xmath3o @xmath13 zeroth moment maps , and 1.3 mm continuum images , of the as 205 system .
note the different grayscales and contour intervals for each row .
based on the continuum measurements , the separation of the two disks is @xmath7 , and the position angle of the binary is 214@xmath15 .
previous works have found separations of @xmath16 and position angles between 204@xmath15 and 213@xmath15 @xcite , with no coherent temporal evolution .
both sources appear to be detected in all three lines and in the continuum , and basic statistics for these detections are listed in table [ table : statistics ] .
for all maps , the synthesized beam is approximately @xmath17 with a position angle of @xmath18 degrees .
the continuum emission appears unresolved or barely resolved , with a spatial fwhm ranging from @xmath19 , and a position angle of 99@xmath15 consistent with that of the synthesized beam .
the line emission is spatially extended in all three isotopologues , and for both components of the system , although the spatial extent of the northern component is larger than for the southern component .
the @xmath1co and @xmath2co emission from the northern component appear asymmetric , with a larger spatial extent towards the north and northwest .
although all isotopologues are present in the zeroth moment maps near as 205 s as well as as 205 n , the @xmath2co emission spatially associated with as 205 s is actually dominated by gas at a similar velocity to as 205 n. therefore , some of this emission may instead be associated with as 205 n. we also note that the c@xmath3o emission near as 205 s is detected in the zeroth moment map only , and at only the 3@xmath20 level .
therefore , we caution that this detection is tentative , and could , for example , arise from errors in continuum subtraction .
channel maps and first moment ( velocity field ) maps for the northern component in @xmath1co and @xmath2co are shown in figures [ fig : channel_plot_12co ] and [ fig : channel_plot_13co ] .
channel maps and first moment maps for the southern component are shown in @xmath1co in figure [ fig : channel_plot_12co_south ] .
line profiles for the three isotopologues are shown in figure [ fig : lineprofile_plot ] .
@xmath1co and @xmath2co lines are computed for the entire system , including the northern and southern components and any extended emission , using a @xmath21 high by @xmath22 wide extraction box . the emission line for c@xmath3o was computed for the northern component only , using a small @xmath23 high by @xmath24 wide extraction box , to reduce the contribution from the noisy background .
the fluxes and linewidths are summarized in table [ table : statistics ] .
the @xmath1co emission has a total flux ( for both components combined ) of 22.25 jy km s@xmath12 , consistent with the flux of @xmath25 jy km s@xmath12 found by @xcite if we allow for a 15% error in absolute flux calibration . using the mean and standard deviation of the line centers
, we estimate @xmath26 of as 205 n to be @xmath27 , consistent to within 1.6 @xmath20 of the published stellar velocity of @xmath28 from @xcite , but inconsistent with the @xmath29 km s@xmath12 found by @xcite .
the high and variable veiling in this source @xcite could contribute to large systematic errors in stellar velocities .
for as 205 s , we take the @xmath26 for @xmath1co ( @xmath30 km s@xmath12 ) to be the most likely value , since we were able to separate the signals from as 205 n and as 205 s ( see table [ table : statistics ] ) .
@xcite find stellar velocities for the components of this spectroscopic binary of [email protected] and @xmath32 km s@xmath12 , with an average of @xmath33 km s@xmath12 , consistent with our result .
llll + @xmath1co & @xmath34 , @xmath35 & @xmath36 , @xmath37 & @xmath38 , @xmath39 + @xmath2co & @xmath40 , @xmath41 & @xmath42 , @xmath43 & @xmath44 , @xmath45 + c@xmath3o & @xmath46 , @xmath47 & @xmath48 & @xmath49 + & flux density ( n , s ) + & [ mjy ] + continuum & @xmath50 , @xmath51 & & + co ( black ) , @xmath2co ( light gray ) and c@xmath3o ( dark gray , and inset ) emission line profiles . @xmath1co and @xmath2co line profiles are derived from the whole system .
the c@xmath3o is from the northern component only , to reduce noise .
excess @xmath1co emission at blue wavelengths is contributed by the southern component .
the vertical dotted line marks the measured cloud velocity near as 205 .
[ fig : lineprofile_plot ] ] to aid with the analysis of the data , we constructed disk models to compare with the observed data .
dust radiative transfer was performed with radmc @xcite , and line radiative transfer with radlite @xcite .
simulated alma datasets were created using the casa routines _ simanalyze _ and _
simobserve_. all simulated observations use the antenna configuration from the observations on 04 may 2012 , and an exposure time of 30 minutes .
simulated datasets are shown in figures [ fig : map_plot ] [ fig : channel_plot_12co_south ] . our nominal disk model for the northern component uses most of the same stellar and disk parameters derived by @xcite from a combined fit to the spectral energy distribution and millimeter visibilities
@xmath52 , @xmath53 and @xmath54 , @xmath55 , @xmath56 and @xmath57 at 80 au , with @xmath58 .
however , in contrast to @xcite who use the similarity solution for surface density , @xmath59 , we use a broken power - law @xmath59 , @xmath60 out to @xmath61 au , and @xmath62 beyond .
we also use a simple grain opacity with 80% small amorphous astronomical silicates and 20% carbon grains .
less information is available in the literature for the southern component , and so for simplicity we utilize many of the same model parameters as for as 205 n. however , we assume @xmath63 , @xmath64 and @xmath65 , values derived by @xcite but with a correction for the fact that the southern source is a spectroscopic binary @xcite . we find a good match to the observed continuum data with a disk mass of @xmath66 .
the co 21 emission is dominated by disk regions with densities well above the critical density for this line ( e.g. * ? ? ?
* ) , and even the more tenuous winds we are investigating here are predicted to have densities at or above the critical density out to large distances from the star @xcite .
therefore , all models assume local thermodynamic equilibrium conditions .
we assumed a gas to dust ratio of 100 , and used a parameterized gas temperature increase in the disk atmosphere , following the procedure described in @xcite .
in particular , we set @xmath67 above the @xmath68 disk surface .
we also assumed an h@xmath69/co ratio of 5000 @xcite and a co freeze - out temperature of 17 k. for isotopologue abundances , we assumed @xmath1co/@xmath2co@xmath70 and @xmath1co / c@xmath3o@xmath71 .
since our limited information would result in strong model degeneracies were we to try to exactly match the observed data , we simply made small adjustments to the total flux produced by the models to approximately match the data .
for as 205 n , we multiplied the continuum models and the @xmath1co , @xmath2co and c@xmath3o continuum - subtracted models by factors of 1.3 , 1.4 , 0.4 and 0.2 . for the southern source models , we multiplied by factors of 1 , 1.2 , 0.5 and 0.3 .
none of the conclusions in this work depend on the absolute flux of the models .
we find that an inclination of @xmath515@xmath72 is required to fit the observed @xmath2co velocity field for as 205 n. previous spectro - astrometric results , which also resolved the binary , found a disk inclination of @xmath520@xmath15 .
using the @xmath1co velocity field , we find a disk inclination of 25@xmath73 for as 205 s. these inclinations and error bars are determined via by - eye examination of the channel maps and residuals , and based on a combination of the relative fluxes in the low and high - velocity channels , as well as the shapes of the maps .
we also find that the disk around as 205 s must be significantly smaller than as 205 n , and set @xmath74 au .
we discuss this result in more detail in section [ section : truncation ] .
there is some uncertainty in the literature about the mass of as 205 s. @xcite derive a larger mass than assumed in our nominal model , @xmath75 , along with temperatures for as 205 n and s of 3800 and 4000 k , and a total luminosity of @xmath76 . with this alternative choice of parameters , we find an inclination for as 205 s of @xmath77 .
given the limited resolution of our observations compared to the size of the as 205 s disk , spatial information can not be used to break the degeneracy between stellar mass and inclination .
figure [ fig : channel_plot_13co ] shows that the as 205 n disk model deviates from the @xmath2co data in some respects . in particular , the high velocity emission is slightly over - predicted .
however , decreasing the inclination of the model would produce maps less elongated than the data at low velocities .
a similar discrepancy was noted by @xcite for tw hya , and attributed to a warp in the tw hya disk . in our case
, the discrepancy could instead be related to the other evidence for non - keplerian motions in our dataset , which is discussed in much greater detail throughout this work .
the position angles of the as 205 n and as 205 s disks are observed to be @xmath78 and @xmath79 , respectively , and are therefore misaligned by nearly 180@xmath15 .
the relative alignments of the disk rotation axes can not be unambiguously determined unless the near and far sides of both disks are known .
the disks are either oriented in the same direction with respect to the plane of the sky , with rotation vectors nearly anti - aligned , or they are oriented in opposite directions with respect to the plane of the sky , with rotation vectors having a misalignment of @xmath80 .
the @xmath2co emission from as 205 n shows a distinct ne / sw asymmetry , with ne regions being brighter .
this is especially apparent in the 4.40 km s@xmath12 channel in figure [ fig : channel_plot_13co ] , for example .
if attributed solely to disk emission , this effect is naturally produced if the ne side is the far side of the disk , due to a combination of projection and optical depth effects ( e.g. * ? ? ?
* ; * ? ? ?
this orientation is consistent with the spectro - astrometric results of @xcite .
however , this interpretation of the asymmetry is not definitive , as our dataset includes several features that are not well explained by a disk model alone .
it is not possible to determine the near / far - side orientation of the disk around as 205 s with these data . in this work ,
our disk models assume a broken power - law surface density that declines sharply at some radius @xmath81 . for as 205 n , the data are consistent with @xmath82 au .
this is demonstrated in figure [ fig : spatial_profile_combined ] , which shows a normalized spatial profile along a cut in the zeroth moment map of @xmath2co emission .
we have chosen to extract an e - w oriented profile to minimize contamination from the extended emission observed towards the nw of the image , as well as from the southern source .
the fwhm of the observed spatial profile is @xmath83 .
we show that the observed @xmath2co spatial profile is well reproduced by our nominal disk model .
the line emission from as 205 s , however , is more compact . because the zeroth moment @xmath1co map is dominated by emission from as 205 n , we consider in this case a single channel map , shown in figure [ fig : spatial_profile_combined ] , which does not suffer any contamination . in this case
, the fwhm of the emission is only @xmath84 and a model with @xmath61 au does not fit the data .
instead , we find that models with @xmath85 au provide a reasonable fit . in principle , the spatial extent of the image also depends on the disk inclination and position angle ; however , we find the spatial extent of the model to be relatively insensitive to position angle and inclinations in the range of @xmath86 .
this truncation can be compared with expectations from binary disk truncation models .
for example , the parametric fit to truncated disk sizes from @xcite predicts a relative size ratio for the two disks of a binary that depends only on the binary mass ratio ( assuming the disks are close enough to each other to be truncated ) . for this system , with @xmath87 and @xmath88 ,
the predicted ratio of disk sizes would be @xmath89 .
our best - fit disk models have ratios near 2 slightly larger than predictions . with the masses assumed by @xcite , @xmath90 and @xmath91 ,
the two disks would be expected to be nearly the same size , which is not consistent with our observations .
co @xmath13 emission from as 205 n. the dashed line shows where the spatial profile is extracted .
b ) the extracted spatial profile for as 205 n derived from the data ( solid line ) and the nominal disk model with @xmath61 au ( dot - dashed ) .
c ) channel map of @xmath1co emission from as 205 s at @xmath92 km s@xmath12 ( same as the 3rd panel in the top half of figure [ fig : channel_plot_12co_south ] ) .
the dashed line shows where the spatial profile is extracted .
d ) the extracted spatial profiles for as 205 s. the solid line shows the data , while the dot - dashed , dashed , and dotted lines show disk models with @xmath93 au , 35 au and 20 au . [
fig : spatial_profile_combined ] ] one of the most striking results of this work is the non - axisymmetric excess emission stretching northeast of as 205 n , seen at the 1020@xmath20 level in the @xmath1co zeroth moment map and channel maps ( figures [ fig : map_plot]a and [ fig : channel_plot_12co ] ) , and at the 3@xmath20 level in the @xmath2co zeroth moment map ( figure [ fig : map_plot]d ) .
the @xmath1co emission , extending to distances of @xmath4 , is much wider than the synthesized beam width with major axis @xmath9 and covers an angular extent of nearly 180@xmath15 .
figure [ fig : channel_plot_12co ] also demonstrates that most of the extended emission consists of low velocity gas within @xmath51.2 km s@xmath12 of the mean source velocity . using the @xmath1co model residuals
, we can estimate the mass and optical depth of the spatially extended gas emission . using the large box shown in figure [ fig : map_plot]c
, we find a total flux @xmath94 jy km s@xmath12 .
we have also computed the flux without including the negative values , and have tested a range of possible scalings for the disk model ( each producing different residuals ) .
these tests imply that the systematic uncertainty in the total flux is a factor of a few .
the temperature of the emission is essentially unknown . assuming the emission is optically thin and has a low excitation temperature of 30 k
, we find a total mass of @xmath1co of @xmath95 , or a total h@xmath69 mass of 1@xmath96 assuming an h@xmath69/co ratio of @xmath97 .
for a much higher excitation temperature of 1000 k , the total mass would be about 20 times larger .
we also compute flux and mass upper limits using the @xmath2co residuals in this same box .
we find a total flux of @xmath98 jy km s@xmath12 ( a 3@xmath20 limit ) and a corresponding mass of @xmath99 of @xmath2co , implying a @xmath1co/@xmath2co ratio @xmath100 .
we can also evaluate the optical depth to determine whether @xmath101 reaches high values .
for the small box in figure [ fig : map_plot]a , assuming the observed flux is proportional to @xmath101 and taking a very conservative ( i.e. narrow , to maximize @xmath101 ) local line broadening of 10 m s@xmath12 , we compute a co column density of @xmath102 g @xmath103 , corresponding to an optical depth of 0.9 .
therefore , the emission may be optically thick , and we consider our mass estimate a lower limit .
the @xmath1co/@xmath2co ratio in this box is @xmath104 , and therefore consistent with either optically thin or moderately optically thick emission , assuming the true isotopic ratio is similar to the values of @xmath105 found in nearby clouds or the interstellar medium @xcite . finally , we can estimate the gas / dust ratio of the extended emission , by comparing the line and continuum emission . in the large box defined in figure [ fig : map_plot]c ,
there is no detected dust emission above the noise level , which has an rms value of @xmath106 mjy beam@xmath12 .
assuming the emission is optically thin , and taking an excitation temperature of 30 k and a grain opacity of @xmath107 @xcite , we find a 3@xmath20 limit to the dust mass of @xmath108 . utilizing the gas mass from above , we find that the gas / dust ratio is @xmath109 . thus , the constraint on the gas / dust ratio of the extended emission is weak , with both sub - canonical ( @xmath110 ) and super - canonical ( @xmath111 ) values permitted by the data .
the @xmath1co and @xmath2co first moment maps also show a distinct `` s''-shaped warp in the velocity field , which is different from expectations for a keplerian disk .
figure [ fig : mom1_comp_map ] shows deviations of the observed velocity fields from our nominal disk model .
the observations deviate from expectation by as much as @xmath51 km s@xmath12 . co velocity fields ( first moment maps ) , with velocity relative to @xmath112 km s@xmath12 .
[ fig : mom1_comp_map ] ]
we have identified low - velocity , wide - angle spatially extended @xmath1co emission from as 205 n , as well as non - keplerian @xmath1co and @xmath2co velocity fields . here
we discuss possible physical explanations for this extended and asymmetric emission .
one possible explanation for our observations is that we are seeing small - scale structure in the surrounding molecular cloud , whose emission is primarily resolved out by the interferometer due to its large spatial scale .
the ophiuchus cloud region was mapped in @xmath1co @xmath113 by @xcite . as 205
n lies to the north of most of the prominent filaments in the region , and so is less likely to suffer from cloud contamination than regions closer to the core of ophiuchus .
more importantly , the local cloud velocity is @xmath53 km s@xmath12 and , as shown in figure [ fig : lineprofile_plot ] , this velocity is far from the observed emission line center .
additionally , much of the spatially extended emission we observe is redder than the emission line center , further separating it from the cloud velocity .
velocities @xmath1144.5 km s@xmath12 are only observed in the @xmath1co @xmath113 cloud maps at large distances ( @xmath115 degrees ) from as 205 n. therefore , we do not believe that molecular cloud emission can be the cause of the observed spatially extended emission . based on a variety of evolutionary indicators , as 205 n is a stage ii source , with a disk but no envelope . according to the classification scheme of @xcite ,
class ii objects ( believed to represent stage ii ) have @xmath116 m spectral indices , @xmath117 , between @xmath118 and @xmath119 , while as 205 n has a 224 @xmath120 m spectral index in the range of @xmath121 to @xmath122 . as 205
n also has a bolometric temperature between 1200 and 1400 k@xmath123 , while class ii objects are defined to have @xmath124 k @xcite .
furthermore , as 205 has no detectable envelope in the millimeter maps obtained by @xcite ( with a beam size of @xmath125 ) , while all of the class 0/i objects in their sample do show resolved envelope emission .
in addition , the total 1.3@xmath0 mm single dish flux density of 450@xmath126 mjy / beam ( assuming a 20% flux uncertainty ; @xcite ) is consistent with our measured system flux density of 441 mjy , suggesting little to no unseen envelope material . attributing the 9 mjy difference to an envelope ( assuming @xmath127 and @xmath128 @xmath129g@xmath12 , which assumes a gas / dust ratio of 100 ) would imply a mass of 0.001 @xmath130 at least an order of magnitude lower than typical envelope masses in ophiuchus @xcite , although the large error bars on the single - dish measurement could allow a mass up to @xmath131 . therefore , as 205 sits in the class ii regime , although the presence of a very low mass envelope can not be excluded .
however , these alma observations represent a new resolution regime for observations of low - j co emission from protoplanetary disks , and studies at similar resolution have so far focused on more evolved transition disks ( e.g. , @xcite , @xcite ) . instead ,
as 205 n s somewhat high spectral index , as well as its relatively high accretion rate ( @xmath132yr@xmath12 ; @xcite ) and disk mass could imply that it sits at the evolutionarily younger end of stage ii .
therefore , in principle this extended emission could represent the first observation of a small - scale , low - mass remnant of the natal envelope
. however , this hypothesis suffers from stability arguments .
cloud material at a distance of @xmath133 should not be stable in a system with a @xmath7 binary ; this distance is too far to be stable circumprimary material , and too close to be stable circumbinary material @xcite .
we consider here whether the observations might be explained as accretion flows from a circumbinary disk .
the arc - like structures seen in the extended @xmath1co emission near as 205 n appear similar to those observed around other @xmath134 separation binaries , such as uy aur , gg tau and sr 24 @xcite , structures which are posited to represent accretion streamers from circumbinary disks .
since this scenario requires the existence of a circumbinary disk , we discuss its possible existence below .
based on the non - detection of a circumbinary ring in @xmath1co emission , as 205 can not have a circumbinary disk similar to those around either gg tau or uy aur .
gg tau has a circumbinary disk extending from @xmath135 from a @xmath136 binary @xcite , with a total mass of @xmath137 @xcite .
the uy aur circumbinary disk has a radius of @xmath5 500 au , and a mass of @xmath138 @xcite . assuming a face - on circumbinary ring extending from @xmath139 ( 250 - 750 au ) , an excitation temperature of 30 k , an h@xmath69:co ratio of 5000 , and using a 3 @xmath20 upper limit derived from our zeroth moment @xmath1co map ( 0.03 jy km s@xmath12 beam@xmath12 )
, we derive an upper limit to the mass of a circumbinary disk around as 205 of @xmath140 significantly lower in mass than that observed around either gg tau or uy aur .
note that we use a co abundance consistent with the @xmath2co abundance utilized by @xcite and so the choice of co abundance does not affect the comparison between the two disks . in addition , we note that our observations are sufficiently sensitive to detect disks with similar surface densities as those observed around gg tau or uy aur
. the circumbinary disk could potentially remain hidden if it were significantly more extended ( and therefore resolved out by the interferometer ) .
we are sensitive to circumbinary structure up to @xmath141 .
since the observed @xmath1co emission extends no more than @xmath4 , it can not represent an accretion streamer from an @xmath142 wide disk .
models of this process ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) show continuity of the accretion streamers out to the inner edge of the circumbinary disk .
it is also possible that significant freeze out of co could hide a circumbinary disk from our view .
however , single - dish flux measurements for this source place upper limits on the total mass of the system . as 205 has a single - dish 1.3 mm flux consistent with its interferometric flux ( as discussed in section [ sec : envelope ] ) , and therefore consistent with little or no unseen material
nevertheless , the large error bars on the absolute value of the 1.3 mm flux could allow for a mass of up to @xmath143 ( where this implicitly assumes a gas / dus ratio of 100 ) .
we note that this situation is in contrast to the binary sr 24 , which shows features in scattered light that have been attributed to accretion streamers @xcite , and , therefore , could be considered an analog to as 205 .
sr 24 has a single - dish flux 34@xmath144 higher than its interferometric flux @xcite , consistent with a significant amount of extended emission a mass of @xmath145 , assuming @xmath146 and @xmath147 k. in conclusion , we show that observations of as 205 place very firm limits on the possible mass of a co - rich disk ( @xmath148 ) , and the single - dish flux is consistent with no unseen dust , although a total disk mass of up to 0.01 @xmath130 ( assuming a gas / dust ratio of 100 ) is allowed by the large error in absolute flux . with no evidence of a large reservoir to supply mass to the circumstellar disks , we believe it s unlikely that we are observing accretion streams from a circumbinary disk .
additional single - dish measurements would be helpful to place firmer limits on any possible unseen material in this system .
another possible explanation for our observations is tidal stripping by as 205 s. this explanation has been previously invoked by @xcite to explain the morphological and kinematic signatures in the disk around the t tauri star rw auriga a , which bears some interesting similarities to our observations .
rw aur is a @xmath149 binary whose @xmath1co 21 emission shows a red - shifted `` tail''-like feature extending to @xmath133 , although at somewhat higher velocities , up to @xmath57 km s@xmath12 equivalent to 22.5 km s@xmath12 if adjusted to the 15@xmath15 inclination of as 205 n. we first evaluate the relevance of tidal effects by computing the tidal radius of the secondary . for @xmath150 and @xmath151
, the binary mass ratio is @xmath152 .
the corresponding lagrangian point @xmath153 @xcite is located at @xmath154 au from @xmath155 , i.e. , reaching just outside the @xmath580 au outer radius of the as 205 n disk . consequently , some perturbations to the disk symmetry are to be expected
. tidal effects will be strongest if the encounter is resonant or quasi - resonant " @xcite . roughly speaking , as in a forced oscillator model , tidal interactions
are most pronounced when there is frequency matching between the internal modes of the victim ( in this case , the disk around as 205 n ) and the external forcing frequency exerted by the perturber , @xmath156 . in other words , maximum response occurs at the disk radii @xmath157 such that the internal orbital frequency @xmath158 is an integer fraction of @xmath159 , where @xmath160 is the pericenter distance , where the resonance condition is given by @xmath161 , and thus the inner lindblad resonance occurs at a radius @xmath162 . ] .
note that the resonance also requires a prograde orbit of the perturber with respect to the victim s spin .
a single passage by the perturber can excite prominent two - armed spirals ( similar in morphology to what we observe ) in the victim provided the resonance occurs within the victim s disk s radial extent ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , if the pericenter distance , @xmath160 , is equal to the binary projected separation of @xmath5160 au or larger , low order near - resonance is not achieved for radii inside the primary s disk with our assumed @xmath163 . to achieve low - order resonance , as 205
s would need to be more massive .
as discussed in section [ sec : models ] , there is some uncertainty about the mass of as 205 s. with the as 205 s mass estimate derived by @xcite , @xmath164 , resonant forcing would occur at @xmath165 au , just inside the outer radius of the as 205 n disk .
however , even if as 205 s is massive enough to have low - order resonant interactions with the disk of as 205 n , a major drawback of this explanation is that bound pairs are expected to be already devoid of gas at the lowest order resonances that create strong two - armed features , since those resonances would be quickly cleared out after a few subsequent pericenter passages @xcite .
therefore , we would need to be observing as 205 during its first , or nearly first , encounter .
this is unlikely if as 205 is a bound pair , as the orbital period is @xmath5 a few thousand years , while the estimated age of as 205 is @xmath166 years @xcite .
it is also very unlikely that this is a first ( flyby ) encounter between two stars .
@xcite suggest that close stellar encounters may be likely during the @xmath56 myr lifetime of disks around stars in massive young clusters ; with encounter probabilities of 10 - 30% for encounter distances of 60200 au ( assuming a cluster mass of @xmath167 and a size of 0.3 pc ) , one can expect @xmath5100 close stellar encounters throughout the lifetime of the cluster . however , ophiuchus is significantly less massive and dense than this model presumes , with a stellar mass of @xmath168 in the central 2 pc@xmath169 of the @xmath170 oph cloud core @xcite , and , furthermore , as 205 lies in a less dense region at the edge of the ophiuchus cloud @xcite . assuming @xmath171 in 2 pc@xmath169 , and using the equations found in @xcite
, we find encounter probabilities of a few @xmath172 out to 300 au equivalent to an expected number of collisions of a few @xmath173 .
finally , we must compute not the likelihood of an encounter , but the likelihood of observing an encounter in progress .
if we assume an encounter velocity of 1 km s@xmath12 and an impact parameter of 500 au , the encounter lasts @xmath52000 years a factor of 500 shorter than the cluster age of at least 1 myr ( e.g. * ? ? ?
* ) ; therefore , the probability of viewing an encounter is 500 times lower than the probability of an encounter occurring .
these factors combined suggest that the probability of observing a flyby encounter with an impact parameter of a few 100 au in this region is @xmath174 .
although lower - order resonances do not appear to explain the features we observe , in the long run , the evolution of circumprimary and circumbinary disks is dominated by higher - order , weaker resonances . as with the lower - order resonances , the resonant forcing and the internal damping of the disk reach a quasi - stationary state , in which the disk size is nearly in equilibrium .
a balance is reached between the viscous and the tidal torques , eventually defining the truncation radius of both circumprimary and circumsecondary disks @xcite . however , one subtlety in the analysis by @xcite is that the equilibrium truncation radius is reached only after a viscous relaxation timescale that depends on the binary eccentricity , the final disk radius , the disk aspect ratio and the disk viscosity .
therefore , for very young systems , transient responses on the disks are still possible during every pericenter passage , especially for very eccentric binaries .
unfortunately , at this point we are not able to pursue this analysis further due to the largely unconstrained orbital properties of this system , and the uncertain mass of as 205 s. our observations might also be explained by some form of wind .
thermal ( photoevaporative ) winds are unlikely to explain our observations , as the driving radiation for these flows tends to destroy molecules ( e.g. * ? ? ?
* ) . on the other hand , magneto - hydrodynamic ( mhd ) disk winds , which are believed to be ( at least partly ) responsible for the larger - scale bipolar jets and outflows emanating from young stars ( e.g. * ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , can produce relatively low - velocity winds with a significant molecular content @xcite .
mhd disk winds have launching velocities that are similar to the sound speed ( e.g. * ? ? ? * ) , and , therefore , tracers of the wind launching region can have relatively low velocities .
temperatures in the launch region range from a few 1000 k in the inner disk , down to @xmath110 k in the outer disk , with associated velocities in the range of @xmath513 km s@xmath12 . observed launching velocities will be even lower if the streamlines are not directed towards the observer .
however , poloidal velocities increase with distance along a streamline , reaching velocities as high as @xmath5 a few 100 km s@xmath12 in existing models @xcite , which are much higher than we observe here .
another obstacle for disk wind models may be whether there is sufficient overall gas density to explain the extended emission . in the models of @xcite ,
co densities of @xmath175 result in a theoretical column density of @xmath176 , assuming a column with vertical height 20 au .
this is three orders of magnitude below the observed lower limit to the column density , @xmath177 .
however , if lower terminal velocities are possible , then mass conservation will correspondingly increase the density in the wind , allowing models to match observations .
we consider this in section [ sec : windmodels ] .
the structure observed here is significantly different from the very collimated millimeter ouflows observed around class 0 and class i disks ( e.g. * ? ? ?
* ) , many of which additionally appear to be tracing shocks on the inner surface of an outflow cavity in a surrounding envelope . however , @xcite imaged a sample of nine class 0 , i and ii young stellar objects in @xmath1co 1 - 0 and noted that the millimeter emission becomes less collimated with evolutionary state .
in fact , the emission we observe bears some resemblance to the structures observed by @xcite near the class ii sources t tau and gk / gi tau .
the size scales of these structure are larger ( up to @xmath178 in length ) , but our observations are only sensitive to structures up to @xmath179 in size .
the velocities of these outflows are also higher than what we observe , typically 25 km s@xmath12 offset from the source velocity . in general , however , millimeter observations of class ii sources with outflows are limited , and even fewer observations have the spatial resolution of our observations , making comparisons difficult . as 205 n has some spectroscopic evidence for an active outflow .
although to our knowledge no optical jets have yet been imaged from this source , @xcite reported strong , narrow , blueshifted ( @xmath180 km s@xmath12 ) absorption features in na d lines that are consistent with outflowing gas and observed in other sources with known outflows . in addition , @xcite report a high - velocity component in observations of the [ o i ] 6300 line from as 205 n , indicative of an outflow . it is also important to note that our observations were _ motivated _ by the suggestion of @xcite that a disk wind may explain observed single - peaked rovibrational co emission lines , as well as asymmetric spectro - astrometric profiles of these same lines . in the next section , we test whether the model utilized by @xcite can also explain our observations . a simple parameterization of a disk wind was implemented by @xcite to fit observed rovibrational emission lines , and incorporated into the radiative transfer code radlite @xcite .
we describe this model and our choice of nominal model parameters in detail in the appendix .
we find that the nominal wind model does not produce any appreciable extended emission , beyond that expected for a disk , due to the low predicted densities at distances of 100 s of au from the central star .
however , the model does reproduce some of the other features of our observations .
as shown in figure [ fig : channel_plot_12co_43_30_54_large ] , it produces additional flux at higher velocities compared to a disk - only model .
the wind model also produces deviations from a keplerian velocity field , including a curved first moment map , demonstrated in figure [ fig : mom1_comp_map_wind ] .
the curvature in the model is due to two effects . at small radii , the dominant effect of the wind
is that it makes the azimuthal velocities sub - keplerian . at larger radii ,
the poloidal velocities are larger and begin to dominate the differences , causing the far side of the disk ( in this case , towards the northeast ) to be preferentially redshifted and the near side ( towards the southwest ) to be preferentially blueshifted .
there are some similarities between the modeled and observed first moment maps , in that both show a complex `` s''-like shape , although it is clear that the observed moment map is more complex than the modeled map .
a twisted velocity field can also arise from radial inflow , as suggested by @xcite to explain the velocity field of the transitional disk around hd 142527 .
inflow and outflow can only be distinguished if the near and far sides of the disks are identified . in detail , hd 142527 is not an analog to as 205 n ; the fast radial inflow in hd 142527 is proposed to reconcile its high accretion rate and inner clearing , while as 205 n has a full disk with no inner clearing .
in addition , radial inflow alone would not explain the very extended @xmath1co emission we observe . however , some wind models themselves predict that accretion may be dominated by restricted regions of the disk producing fast flows @xcite
. the resultant signature in the velocity field would then depend on the relative contributions of inflowing and outflowing gas to the observed emission , and could vary throughout the disk , creating a complex velocity field .
unfortunately , disk - wide simulations of this process do not yet exist for comparison with observations .
co first moment maps , relative to @xmath112 km s@xmath12 , as well as their difference .
bottom : disk+wind and disk - only models as well as their difference .
all models assume the far side of the disk is located towards the northeast .
[ fig : mom1_comp_map_wind ] ] although the _ nominal _ wind model does not produce any appreciable extended emission , the density of the wind depends on the assumed mass - loss rate and terminal velocity .
figure [ fig : channel_plot_12co_43_30_54_large ] shows predicted channel maps for an alternative wind model , with a low terminal velocity ( 2 km s@xmath12 ) and high mass - loss rate ( @xmath181 yr@xmath12 ) , which shows appreciable extended emission .
the detailed structure does not obviously match our observations .
it should be noted , however , that dynamical timescales for the binary orbit and the wind are similar , and so as 205 s should cause perturbations to the wind structure . therefore , our observations may be explained by a _
combination _ of disk winds and binary interaction . in conclusion ,
wind models are in some ways consistent with the observations of as 205 n. in particular , models produce a curved first moment map with some similarities with the observed @xmath1co velocity field . however
, existing wind models do not readily produce significant extended emission .
extended emission instead requires that the models be modified to have lower terminal velocities and/or higher mass - loss rates .
although the geometry of the resulting emission does not precisely match our observations , the detailed features we observe can not be produced by an axisymmetric model .
they could , however , be produced by interactions between the wind and the stellar companion , as 205 s.
these are the highest resolution observations of rotational co emission from as 205 to date . as 205 was observed in @xmath1co @xmath182 by @xcite with a beam size of @xmath183 and in @xmath1co @xmath13 by @xcite with a beam size of @xmath184 . in both cases ,
the beam size was not sufficient even to resolve the binary . since the extended structure we observe is only a few arcsec in size , a resolution of @xmath134 or better
is required .
only a few disks have been observed in molecular lines at a resolution of @xmath134 or better to date , as pre - alma observations were limited to the brightest disks .
therefore , it is difficult to say whether the phenomenon we observe is a common or unusual feature for protoplanetary disks . interestingly
, molecular line imaging at or near this resolution often seems to reveal surprises for example , an extended co disk around tw hya @xcite , an asymmetric , perhaps tidally - induced `` arm '' near rw aur @xcite and asymmetric spiral arms in the disk of ab aur @xcite .
alma observations of a few disks have already revealed azimuthal asymmetries and non - keplerian flows @xcite , as well as a helical disk wind @xcite .
therefore , it seems likely that the resolution and sensitivity afforded by alma will continue to reveal many new small - scale asymmetric structures .
if the extended structure we observe here is a common feature for protoplanetary disks , this could have implications for the interpretation of lower - resolution observations . for low resolution observations ,
the observed co flux could incorrectly be attributed solely to a disk .
although it s difficult to determine the precise contribution to the flux from the wind , the box shown in figure [ fig : map_plot]c contributes 5% of the total observed @xmath1co @xmath13 line flux .
also , if the wind has a different temperature than the disk , interpretation of observations of multiple co isotopologues could incorrectly determine the gas temperatures .
since gas temperatures are in turn used to determine gas masses ( e.g. * ? ? ?
* ) , this could have important consequences for our understanding of the planet - forming capabilities of disks .
if we are actually observing a wind , with some simple assumptions , we can estimate the mass - loss rate . for this calculation
, we use the co observed in the large box in figure [ fig : map_plot]c ( see section [ sec : mass ] ) and multiply by 4 to account for all directions of the wind .
we obtain @xmath185 for an excitation temperature of 30 k , or @xmath186 for 1000 k. the mass - loss rate is @xmath187 where @xmath188 and @xmath189 are the actual velocity and length of the outflow . both @xmath189 and @xmath188 are different from the observed distance and velocity , where @xmath190 , @xmath191 and @xmath192 is the angle between the outflow and the observer .
thus , @xmath193 .
taking @xmath194 au , @xmath195 km s@xmath12 and assuming @xmath196 ( equal to the best - fit disk inclination ) , we find @xmath197 .
a different assumption about the inclination of the flow can either raise ( for @xmath198 ) or lower ( for @xmath199 ) the mass - loss rate .
the disk + wind models described in section [ sec : windmodels ] also required a rate of @xmath200 to produce significant extended emission . using observations of co rovibrational lineshapes and spectro - astrometry along with disk+wind radiative transfer models
, @xcite estimated a similar or slightly lower mass - loss rate for as 205 n @xmath201 .
both results therefore suggest that @xmath202 .
we have presented here one of the most sensitive and highest resolution mm - wave studies of a nearby t tauri binary system and its accompanying gas disks , a study which was enabled by the newly available alma facility .
it is clear from these observations that as 205 n does not solely consist of a keplerian gas disk , with differences including extended asymmetric @xmath1co emission and a non - keplerian velocity field .
although we have examined here several possible scenarios to explain these observations , no single scenario is able to completely explain these features .
two of the most promising explanations include a low - velocity wind , or tidal stripping by the companion , but potential models will need to be explored further and fine - tuned if they are to match the observations
. it may also be that we are observing wind - driven mass - loss , modified by tidal interactions with the secondary .
future alma observations of disks in close binary systems as well as disks with ir spectroscopic evidence for disk winds , or evidence for larger - scale outflows , will be helpful for distinguishing between the two possible explanations . aside from the observations presented here , it is known that as 205 n has other somewhat unusual properties compared to typical t tauri disks
. however , it is _ not _ unique , as there appears to be an emerging sub - population of disks sharing properties with as 205 n , properties including high accretion rates and veiling , as well as strongly single - peaked ir emission line profiles ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the story that these disks have to tell us is still unclear , but we are excited about the prospects of continuing to explore their interesting properties , and expect that the mysteries of their origin will soon be revealed .
we would like to thank the staff at the north american alma science center , especially scott schnee , for their help with the data reduction .
c.s . would like to acknowledge helpful discussions with ilaria pascucci , sean andrews and katherine rosenfeld .
c.s . would also like to acknowledge the financial support of the noao leo goldberg fellowship program .
d.m . would like to acknowledge helpful discussions with dong lai and robert harris .
this paper makes use of the following alma data : ads / jao.alma#2011.0.00531.s .
alma is a partnership of eso ( representing its member states ) , nsf ( usa ) and nins ( japan ) , together with nrc ( canada ) and nsc and asiaa ( taiwan ) , in cooperation with the republic of chile .
the joint alma observatory is operated by eso , aui / nrao and naoj .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
the wind model used in this work is based on the split - monopole parameterization described by @xcite , and originally from @xcite , in which the poloidal velocities follow a @xmath203-velocity law , the toroidal velocities conserve angular momentum , and the density is determined via mass conservation .
the equations for velocity and density as a function of cylindrical radius , @xmath204 , and distance along the streamline , @xmath205 , are given in @xcite .
the geometry of this model is completely defined by the ( virtual ) launching point , which is located a distance , @xmath206 , below the star , and inner and outer radial extents , @xmath207 and @xmath208 .
this parameterized model is meant to approximate magnetocentrifugal wind models , originally suggested by @xcite and @xcite .
the wind model was implemented into the line radiative transfer code radlite by @xcite .
radlite takes as input an arbitrary density , velocity and temperature grid . in this implementation
the densities and velocities above some defined surface level , @xmath209 , are simply replaced by the equations found in @xcite .
@xmath209 is a vertical optical depth computed at 0.55 @xmath120 m . the temperatures , and the densities below @xmath209 , take whatever form the user may choose ;
in our case , they are described by the disk model discussed in section [ sec : models ] .
parameters used in our nominal model are similar to those used by @xcite and are listed in table [ table : parameters ]
. short definitions of the parameters are provided ; detailed information about these parameters can be found in @xcite .
we also note a few clarifications . in our implementation ,
we replace the product @xmath210 , the product of the local escape velocity and the ratio of the terminal velocity to the escape velocity , with a single parameter , @xmath211 , representing the terminal velocity of the wind .
although @xmath212 varies with radius , in practice the small launching region of the wind precludes a large range in @xmath212 , so this substitution is a useful conceptual simplification .
in addition , we provide the total mass - loss rate ( @xmath213 ) , integrated over the disk surface participating in the wind , although we also include the mass - loss rate per unit area at the innermost launching point ( @xmath214 ) defined in @xcite .
lll d & 18.5 @xmath215 & 1 + @xmath203&2&2 + @xmath211&20 km s@xmath12&3 + @xmath216 & @xmath217&4 + @xmath218 & @xmath219&5 + @xmath160 & -7/2&6 + @xmath220&19 au&7 + @xmath209&0.75&8 + @xmath207&0.14 au&9 + @xmath208&1 au&10 + alencar , s. h. p. , johns - krull , c. m. , & basri , g. 2001 , , 122 , 3335 andrews , s. m. , & williams , j. p. 2005
, , 619 , l175 andrews , s. m. , & williams , j. p. 2007
, , 671 , 1800 andrews , s. m. , & williams , j. p. 2007
, , 659 , 705 andrews , s. m. , wilner , d. j. , hughes , a. m. , qi , c. , & dullemond , c. p. 2009
, , 700 , 1502 andrews , s. m. , wilner , d. j. , hughes , a. m. , qi , c. , & dullemond , c. p. 2010
, , 723 , 1241 andrews , s. m. , wilner , d. j. , hughes , a. m. , et al .
2012 , , 744 , 162 andre , p. , & montmerle , t. 1994 , , 420 , 837 arce , h. g. , & sargent , a. i. 2006 , , 646 , 1070 artymowicz , p. , & lubow , s. h. 1994 , , 421 , 651 artymowicz , p. , & lubow , s. h. 1996 , , 467 , l77 bai , x .-
n . , & stone , j. m. 2013 , , 769 , 76 barnes , j. e. , & hernquist , l. 1996 , , 471 , 115 bast , j. e. , brown , j. m. , herczeg , g. j. , van dishoeck , e. f. , & pontoppidan , k. m. 2011 , , 527 , a119 beckwith , s. v. w. , sargent , a. i. , chini , r. s. , & guesten , r. 1990 , , 99 , 924 blandford , r. d. , & payne , d. g. 1982 , , 199 , 883 brown , j. m. , pontoppidan , k. m. , van dishoeck , e. f. , et al .
2013 , , 770 , 94 bruderer , s. , van dishoeck , e. f. , doty , s. d. , & herczeg , g. j. 2012 , , 541 , a91 bruderer , s. , van der marel , n. , van dishoeck , e. f. , & van kempen , t. a. 2014 , , 562 , a26 cabrit , s. , pety , j. , pesenti , n. , & dougados , c. 2006 , , 452 , 897 casassus , s. , van der plas , g. , m , s. p. , et al .
2013 , , 493 , 191 chen , h. , myers , p. c. , ladd , e. f. , & wood , d. o. s. 1995 , , 445 , 377 christiaens , v. , casassus , s. , perez , s. , van der plas , g. , & mnard , f. 2014 , , 785 , l12 close , l. m. , dutrey , a. , roddier , f. , et al .
1998 , , 499 , 883 corder , s. , eisner , j. , & sargent , a. 2005 , , 622 , l133 donghia , e. , vogelsberger , m. , faucher - giguere , c .- a . , & hernquist , l. 2010 , , 725 , 353 dullemond , c. p. , & dominik , c. 2004 , , 417 , 159 dutrey , a. , guilloteau , s. , & guelin , m. 1997 , , 317 , l55 duvert , g. , dutrey , a. , guilloteau , s. , et al .
1998 , , 332 , 867 eisner , j. a. , hillenbrand , l. a. , white , r. j. , akeson , r. l. , & sargent , a. i. 2005 , , 623 , 952 ferreira , j. , dougados , c. , & cabrit , s. 2006 , , 453 , 785 de geus , e. j. , bronfman , l. , & thaddeus , p. 1990
, , 231 , 137 gahm , g. f. , walter , f. m. , stempels , h. c. , petrov , p. p. , & herczeg , g. j. 2008 , , 482 , l35 ghez , a. m. , neugebauer , g. , & matthews , k. 1993 , , 106 , 2005 gorti , u. , & hollenbach , d. 2009 , , 690 , 1539 goto , m. , usuda , t. , takato , n. , et al .
2003 , , 598 , 1038 greene , t. p. , wilking , b. a. , andre , p. , young , e. t. , & lada , c. j. 1994 , , 434 , 614 gnther , r. , & kley , w. 2002 , , 387 , 550 hanawa , t. , ochi , y. , & ando , k. 2010 , , 708 , 485 herczeg , g. j. , walter , f. m. , linsky , j. l. , et al .
2005 , , 129 , 2777 jrgensen , j. k. , bourke , t. l. , myers , p. c. , et al .
2007 , , 659 , 479 kenyon , s. j. , lada , e. a. , & barsony , m. 1998 , , 115 , 252 klaassen , p. d. , juhasz , a. , mathews , g. s. , et al .
2013 , , 555 , a73 knigge , c. , woods , j. a. , & drew , j. e. 1995 , , 273 , 225 koresko , c. d. 2002 , , 124 , 1082 kurosawa , r. , harries , t. j. , & symington , n. h. 2006 , , 370 , 580 lacy , j. h. , knacke , r. , geballe , t. r. , & tokunaga , a. t. 1994 , , 428 , l69 liu , m. c. , graham , j. r. , ghez , a. m. , et al .
1996 , , 461 , 334 loinard , l. , torres , r. m. , mioduszewski , a. j. , & rodrguez , l. f. 2008 , , 675 , l29 mamajek , e. e. 2008 , astronomische nachrichten , 329 , 10 mayama , s. , tamura , m. , hanawa , t. , et al .
2010 , science , 327 , 306 mccabe , c. , ghez , a. m. , prato , l. , et al .
2006 , , 636 , 932 mcmullin , j. p. , waters , b. , schiebel , d. , young , w. , & golap , k. 2007 , astronomical data analysis software and systems xvi , 376 , 127 melo , c. h. f. 2003 , , 410 , 269 mihos , j. c. , & hernquist , l. 1996 , , 464 , 641 mundt , r. 1984 , , 280 , 749 munoz , d. & kratter , k. , in prep murray , c. d. , & dermott , s. f. 1999 , solar system dynamics by murray , c. d. , 1999 , najita , j. , carr , j. s. , & mathieu , r. d. 2003 , , 589 , 931 natta , a. , testi , l. , & randich , s. 2006 , , 452 , 245 berg , k. i. , qi , c. , fogel , j. k. j. , et al .
2010 , , 720 , 480 berg , k. i. , qi , c. , fogel , j. k. j. , et al .
2011 , , 734 , 98 ochi , y. , sugimoto , k. , & hanawa , t. 2005 , , 623 , 922 paczynski , b. 1977 , , 216 , 822 pani , o. , hogerheijde , m. r. , wilner , d. , & qi , c. 2008 , , 491 , 219 panoglou , d. , cabrit , s. , pineau des forts , g. , et al .
2012 , , 538 , a2 pichardo , b. , sparke , l. s. , & aguilar , l. a. 2005 , , 359 , 521 pitu , v. , guilloteau , s. , & dutrey , a. 2005 , , 443 , 945 pitu , v. , dutrey , a. , & guilloteau , s. 2007 , , 467 , 163 pitu , v. , gueth , f. , hily - blant , p. , schuster , k .- f . , & pety , j. 2011 , , 528 , a81 pontoppidan , k. m. , blake , g. a. , & smette , a. 2011 , , 733 , 84 pontoppidan , k. m. , meijerink , r. , dullemond , c. p. , & blake , g. a. 2009 , , 704 , 1482 pontoppidan , k. m. , salyk , c. , blake , g. a. , et al .
2010 , , 720 , 887 prato , l. , greene , t. p. , & simon , m. 2003 , , 584 , 853 pudritz , r. e. , & norman , c. a. 1983 , , 274 , 677 pudritz , r. e. , ouyed , r. , fendt , c. , & brandenburg , a. 2007 , protostars and planets v , 277 ray , t. , dougados , c. , bacciotti , f. , eislffel , j. , & chrysostomou , a. 2007 , protostars and planets v , 231 reipurth , b. , & zinnecker , h. 1993 , , 278 , 81 rigliaco , e. , pascucci , i. , gorti , u. , edwards , s. , & hollenbach , d. 2013 , , 772 , 60 roddier , c. , roddier , f. , northcott , m. j. , graves , j. e. , & jim , k. 1996 , , 463 , 326 rosenfeld , k. a. , qi , c. , andrews , s. m. , et al .
2012 , , 757 , 129 rosenfeld , k. a. , andrews , s. m. , hughes , a. m. , wilner , d. j. , & qi , c. 2013 , , 774 , 16 rosenfeld , k. a. , chiang , e. , & andrews , s. m. 2014 , , 782 , 62 salyk , c. , herczeg , g. j. , brown , j. m. , et al .
2013 , , 769 , 21 simon , m. , dutrey , a. , & guilloteau , s. 2000 , , 545 , 1034 salyk , c. , pontoppidan , k. m. , blake , g. a. , et al .
2008 , , 676 , l49 thies , i. , kroupa , p. , & theis , c. 2005 , , 364 , 961 wang , h. , mundt , r. , henning , t. , & apai , d. 2004 , , 617 , 1191 white , r. j. , & ghez , a. m. 2001 , , 556 , 265 wilking , b. a. , meyer , m. r. , robinson , j. g. , & greene , t. p. 2005
, , 130 , 1733 yang , h. , herczeg , g. j. , linsky , j. l. , et al .
2012 , , 744 , 121 zhang , k. , pontoppidan , k. m. , salyk , c. , & blake , g. a. 2013 , , 766 , 82 | in this study , we present high - resolution millimeter observations of the dust and gas disk of the t tauri star as 205 n and its companion , as 205 s , obtained with the atacama large millimeter array . the gas disk around as 205 n , for which infrared emission spectroscopy demonstrates significant deviations from keplerian motion that has been interpreted as evidence for a disk wind @xcite , also displays significant deviations from keplerian disk emission in the observations presented here .
detections near both as 205 n and s are obtained in 1.3@xmath0 mm continuum , @xmath1co 21 , @xmath2co 21 and c@xmath3o 21 .
the @xmath1co emission is extended up to @xmath4 from as 205n , and both @xmath1co and @xmath2co display deviations from keplerian rotation at all angular scales .
two possible explanations for these observations hold up best to close scrutiny tidal interaction with as 205 s or disk winds ( or a combination of the two ) , and we discuss these possibilities in some detail . |
strong gravitational lensing due to the dense cores of clusters of galaxies leads to the formation of multiple images and giant luminous arcs of background galaxies .
the lensing observables such as image positions and flux - ratios in turn provide strong constraints on the underlying gravitational potential of the lensing clusters .
the strongly lensed images of background galaxies have therefore been powerful observational tools to probe the mass distribution in cluster cores ( mellier et al .
1993 ; kneib et al .
1993 , 1996 , 2003 ; hattori , makino , & kneib 1999 ; molikawa & hattori 1999 ; broadhurst et al .
2005 ) . exploring high-@xmath3 galaxies is extremely important for understanding unsolved physical properties of forming galaxies as well as understanding the cosmic reionization history .
the bottleneck is that such high-@xmath3 forming galaxies are too faint to be observed with present - day telescopes due to the @xmath4-dependence of the surface - brightness dimming effect .
the most promising method is to utilize the gravitational magnification by clusters as a biggest natural telescope gravitational lens telescope ( ellis et al .
2001 ; hu et al . 2002
; kneib et al .
furthermore , detailed study of high redshift galaxies can only be made for strongly lensed hence magnified galaxies ( e.g. , pettini et al .
2002 on cb58 ) . we have discovered a high-@xmath3 lensed galaxy during the course of spectroscopic follow - up of the rxj0152.7@xmath51357 cluster at @xmath1 as a part of our distant cluster project on the subaru called pisces ( kodama et al . 2005 ) . in this paper ,
we present the spectroscopic and photometric properties of the newly discovered triple lensed images based on the observational data taken with the ground - based subaru telescope as well as on the data with the high - resolution archival advanced camera for surveys ( acs ) on the hubble space telescope ( hst ) . a gravitational lens model of the cluster mass distribution is developed for the three lensed images .
throughout the paper we assume a @xmath6-dominated flat cosmology with @xmath7 , and @xmath8 , and we use magnitudes given in the ab system .
[ fig : spec ] ( 80mm,80mm)fig1.eps the lensed objects were first found as those having very similar colors in @xmath9 based on the optical imaging with suprime - cam on the subaru telescope ( see kodama et al .
2005 for details of the optical data ) .
the photometric properties of the objects are summarized in table 1 .
all the three objects show strong continuum breaks in the @xmath10 color ( @xmath11 with a relatively flat sed at longer wavelength regime .
these photometric properties suggest that their redshifts are high ( @xmath12 ) in analogy to lyman break galaxies .
we have conducted spectroscopic follow - up observations during 1114 october 2004 with focas ( kashikawa et al .
2002 ) in the mos mode .
we used a 300 lines @xmath13 grating blazed at 5500 @xmath14 with the order - cut filter sy47 .
the wavelength coverage was between 4700@xmath15 and 9400 @xmath15 with a pixel resolution of @xmath16 .
a slit width was set to @xmath17 , which gave a resolution of @xmath18 .
we obtained four 1800s exposures for each object nodded by @xmath19 for each shot .
the object @xmath20 was slipped out of the slit for one shot due to the nodding , and the net exposure on this object is @xmath21s .
data reduction was performed in a standard manner using _ iraf_. the objects @xmath22 and @xmath20 were very close to the spatial edge of the slits , and hence we will not discuss relative fluxes of these objects from the spectra .
figure 1 shows the obtained spectra of the the three images under concern .
all the three show a strong emission line around @xmath23 .
the emission line profile for each object was fitted with a gaussian .
the central wavelengths of the lines are 5992.41@xmath15 , 5991.95@xmath15 , and 5992.11@xmath15 for @xmath22 , @xmath20 , and @xmath24 , respectively , with a statistical error of @xmath25 .
therefore , the objects lie at the same redshift within the errors . the objects @xmath22 and @xmath24 show a clear continuum break shortward of this line , although no such break is seen in @xmath20 due to the low _
s / n_. no other lines are observed in the spectra .
these objects are either ( 1 ) a gravitationally lensed ly@xmath26 emitter at @xmath27 or ( 2 ) foreground [ oii ] emitters at @xmath28 .
the latter possibility is rejected , however .
we observe no [ oiii ] lines at the expected wavelength as shown in fig .
it is highly unlikely that galaxies show such a strong [ oii ] emission without any detectable [ oiii ] emission .
the photometric properties of the objects are inconsistent with a @xmath28 galaxy either .
their @xmath10 and @xmath29 colors are too red for their @xmath30 colors to be star forming galaxies at @xmath31 .
we therefore conclude that these objects are ly@xmath26 emitters at @xmath27 .
note that the rms error in the redshifts is @xmath32 . in summary , these objects show very similar photometric properties and they lie at exactly the same redshift these facts strongly suggest that they are originally a single galaxy strongly lensed by the foreground cluster at @xmath1 .
we also obtained the `` on - the - fly '' processed hst / acs images of the field from the hst archive ( hst gto proposal 9290 : ford , h. ) .
the cluster was observed in a @xmath33 mosaic pattern with an overlap of @xmath34 regions in f675w , f775w , and f850lp filters .
the total exposure per pointing is about 4.8 ks for each filter .
we show in fig .
2 the hst image of the lensed objects and the subaru image in fig . 3 .
the objects all apparently look like the edge - on disk , which is probably due to the effect of strong lensing .
[ tab : so ] lccccccccc object & r.a . & dec .
& @xmath35 & @xmath10 & @xmath29 & @xmath30 & @xmath36 + & ( j2000 ) & ( j2000 ) & ( ab - mag ) & ( ab - mag ) & ( ab - mag ) & ( ab - mag ) & ( @xmath37 ) + @xmath22 & 01 52 45.32 & -13 57 08.6 & @xmath38 & @xmath39 & @xmath40 & @xmath41 & 0.141 + @xmath20 & 01 52 45.14 & -13 57 04.9 & @xmath42 & @xmath43 & @xmath44 & @xmath45 & 0.125 + @xmath24 &
01 52 44.41 & -13 56 56.7 & @xmath46 & @xmath47 & @xmath48 & @xmath49 & 0.127 + + + + [ tab : sources ] [ fig : image ] ( 80mm,80mm)fig2.eps
[ fig : gl ] ( 85mm,85mm)fig3.eps the rxj0152 gravitational lens system shows the colinear , three multiple images @xmath50 of a source galaxy at @xmath0 , from which one expects that the lensed images could be associated with a cusp caustic produced by an elliptical potential ( e.g. , schneider , ehlers , & falco 1992 ) . in constructing a lens model of the cluster mass distribution , we closely follow weak lensing mass estimates by jee et al .
( 2005 ) based on the hst / acs observations with superb angular - resolution and photometric - depth .
we adopt jee et al.s parameterized singular isothermal ellipsoid model ( sie : kormann et al .
1994 ; king & schneider 2001 ) centered between the two central brightest cluster galaxies , for describing the elongated global mass distribution .
the sie requires three model parameters : einstein radius @xmath51 , minor - to - major axis ratio @xmath52 , and orientation angle @xmath26 .
jee et al .
obtained the best - fit parameters of @xmath53 ( @xmath54 ) and @xmath55 .
this einstein angle corresponds to the 1-d velocity dispersion of @xmath56 . for the source redshift of @xmath57 , this angle
is scaled to @xmath58 .
the cluster contains several mass clumps associated with the cluster galaxy concentrations . as the three images are located in between the cluster center and the north - east substructure ( mass clump b of jee et al . )
, we take this clump into account in our lens model in the following way : we assume the clump mass of @xmath59 estimated by jee et al .
is the virial mass of the clump .
we then convert the mass into the 1-d velocity dispersion of an sis halo ( @xmath60 : see bartelmann , king , & schneider 2001 ) .
accordingly , our model consists of one sie as the global cluster halo and one sis as the north - east substructure .
we treat the sie parameters ( @xmath51 , @xmath61 , @xmath26 ) and the ( un - lensed ) source position @xmath62 as free parameters of our lens model .
we then constrain the 5 model parameters via @xmath63-fitting to the observation data : the positions and the fluxes of the three images . in the model - fitting
, we use as constraints the thee image positions @xmath64 relative to the center of the sie halo ; we further include the flux ratio @xmath65 of the closer pair as it gives a strong constraint on the location of the critical curve . the @xmath63 is then given as @xmath66 ^ 2 } { \sigma_{i_2/i_1}^2 } , \ ] ] where quantities with superscript m denote the model predictions , @xmath67 is the positional uncertainty of the multiple images , @xmath68 , and @xmath69 is the rms error of the flux ratio @xmath65 .
the image positions @xmath70 of a source with the position @xmath62 are obtained by solving the lens equation @xmath71 ( schneider et al . 1992 ) with the effective lensing potential @xmath72 in sum of an sie and an sis potential .
we show in fig .
3 the derived best - fit model of the rxj0152 lens system overlaied on the subaru image .
the minimum @xmath63 is @xmath73 for @xmath74 degrees of freedom ( dof ) .
we quote the best - fit sie parameters are @xmath75 , @xmath76 , and @xmath77 measured east of north , with the @xmath78 ( @xmath79 ) errors estimated from @xmath80 ( 4 ) in the 5 parameter space .
the predicted source position is inside and close to a cusp caustic produced by the main cluster potential as expected .
for such an image configuration , the image @xmath20 is predicted to have opposite parity as @xmath22 and @xmath24 . in particular , the closest images @xmath81 are mirror images of each other with respect to the critical curve ( see fig .
the high - resolution acs images reveal more detailed structures of the lensed images that were not spatially resolved in the subaru images .
figure 2 clearly shows that the observed image parities are consistent with the lensing hypothesis : each image contains a bright core and extended emission with lower surface brightness , and a clear mirror symmetry between @xmath22 and @xmath20 is seen in the hst / acs data .
the predicted amplifications using the best - fit model are @xmath82 , @xmath83 , and @xmath84 for @xmath85 , and @xmath24 , respectively .
the total amplification factor over the three images is hence @xmath86 .
accordingly , the predicted magnification ratios are @xmath87 and @xmath88 , whereas the flux ratios from the subaru @xmath89-band data are @xmath90 and @xmath91 .
our simple lens model thus reproduces correctly the observed flux ratio of @xmath65 while it fails for @xmath92 ; more than a factor 2 smaller than the observed flux ratio . in this study
, we will not assess the significance of the difference between the observed and predicted flux - ratios for @xmath92 as it requires more detailed lens modeling for constraining all possible substructures ( cf .
kneib et al .
it is expected that the image @xmath24 could be easily magnified by local perturbers since its angular size ( see table 1 ) is sufficiently small compared with that of the lensing potential of possible perturbers ( chiba 2002 ) .
the derived lens model allows us to inspect the intrinsic properties of the un - lensed source from the observed images . in what follows
, we only refer to the images @xmath22 and @xmath20 for which our lens model can accurately reproduce the observations . from the predicted magnifications ,
the magnitude of the un - lensed source is estimated as @xmath93 .
this corresponds to a uv - luminosity of @xmath94 erg / s/@xmath14 or @xmath95 mag . from this
, we estimate a star formation rate ( sfr ) of the source galaxy to be @xmath96 without dust extinction correction ( madau et al .
the observed ly@xmath26 flux indicates sfr@xmath97 ( kennicutt 1998 ; brocklehurst 1971 ) , which is lower than sfr(1560@xmath14 ) .
this is suggestive of dust extinction .
further , the superb hst / acs angular resolution together with our lens model gives us size information of the source galaxy .
the half - light radii @xmath98 of the lensed images are listed in table 1 ( flux_radius given by sextractor : bertin & arnouts 1996 ) .
note that even for the acs images the fwhm of psf is comparable to that of the lensed images . taking this into account
, we give a constraint on the intrinsic source size as @xmath99 ( @xmath100 ; see fig .
2 ) , or @xmath101 pc ( @xmath102 pc ) in physical scales , assuming a simple one - dimensional amplification near the critical curve ( see ellis et al .
2001 for a similar case ) .
we report the discovery and spectroscopic confirmation of three gravitationally lensed images of a distant galaxy in the background of the rich cluster rxj0152.7@xmath51357 ( @xmath1 ) .
the similarity of the photometric and spectroscopic properties of the three images from subaru observations provides strong evidence that they are gravitationally lensed images of a ly@xmath26 emitter at @xmath0 .
a lens model was constructed on the basis of jee et al.s weak lensing mass estimates from the hst / acs data .
our simple lens model can account simultaneously for the observed image configuration and the flux ratio of the closer pair located close to the critical curve .
the image parities in the acs image are also consistent with the lensing hypothesis . to further investigate the properties of this source galaxy at @xmath0 , we need longer wavelength data , such as @xmath103-band which samples its rest - frame @xmath104-band light or _
spitzer / astro - f _ bands at 38@xmath105 m which sample the light from underground old populations related to the stellar mass of this galaxy .
a more detailed modeling of the lens mass distribution is important , as a well - constrained cluster mass model in conjunction with different sets of multiple images with suitable redshifts serves as a powerful probe of the geometry of the universe ( futamase & yoshida 2001 ; golse , kneib , & soucail 2002 ) .
we acknowledge m. chiba , m. takada , t. chiueh , and b .- c .
hsieh for useful discussions .
we thank our referee , j .-
kneib , for his useful remarks helping to improve the paper .
this work is financially supported in part by a grant - in - aid for the scientific research ( no .
15740126 ) by the ministry of education , culture , sports , science and technology .
m.t . acknowledges the jsps research fellowship .
some of the data presented in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 . | we report the discovery and spectroscopic confirmation of three gravitationally - lensed images of a galaxy at @xmath0 in the background of a distant , rich cluster of galaxies at @xmath1 , on the basis of observations with faint object camera and spectrograph ( focas ) on the subaru telescope .
we construct a simple lens model of the cluster mass distribution based on jee et al.s weak lensing mass estimates from deep , high - resolution images by advanced camera for surveys ( acs ) on the hubble space telescope .
the lens model can account simultaneously for the observed image configuration and the flux ratio of the closer pair located close to the critical curve .
the parities of the three images are also consistent with the lensing hypothesis .
since this galaxy is apparently bright ( @xmath2 ) for its redshift due to the magnification , it serves as a good high redshift target on which we can make extensive and detailed studies based on multi - wavelength imaging and spectroscopy . |
quantum information science promises exponential improvement and new functionality for particular tasks in computation @xcite , metrology @xcite , lithography @xcite and communication @xcite .
photonics appears destined for a central role owing to the wide compatibility , low - noise and high - speed transmission properties of photons @xcite
. however , future quantum technologies and fundamental science will require integrated optical circuits that offer high - fidelity and stability whilst enabling scalability . silica - on - silicon waveguide photonic quantum circuits @xcite are an important step , however , conventional lithography is costly , time consuming , and limited to two - dimensions .
here we demonstrate an alternative fabrication technique based on ultrafast laser processing @xcite that overcomes all of these issues .
quantum technologies rely on transmitting and processing information encoded in physical systems typically two - state _ qubits_exhibiting uniquely quantum mechanical properties @xcite .
photons hold great promise as qubits given their light - speed transmission , ease of manipulation at the single qubit level , low noise ( or _ decoherence _ ) properties and the multiple degrees of freedom available for encoding qubits or higher - level systems ( including polarization , optical mode , path and time - bin ) .
the problem of realizing interactions between single photonic qubits was theoretically solved in a breakthrough scheme for implementing optical non - linear interactions using only single photon sources , linear optical elements and single photon detectors @xcite .
remarkable progress in the development of single photon sources @xcite and detectors @xcite , make a photonic approach to quantum technologies very promising . indeed , there have been a number of important proof - of - principal demonstrations of quantum optical circuits for communication @xcite , computing @xcite , metrology @xcite , and lithography @xcite .
however , the use of large - scale ( bulk ) optics to create these circuits places extremely stringent requirements on the alignment and stability in position of the optical components , thus making such an approach inherently unscalable . successful operation of a quantum optical circuit requires that the individual photons be brought together at precisely the same position and with the correct phase on a succession of optical components in order to realize the high - fidelity classical and quantum interference that lies at the heart of single photon interactions @xcite .
integrated optical quantum circuits based on chip - scale waveguide networks will likely find important applications in future quantum information science along side optical fibre photonic quantum circuits that have been demonstrated in quantum key distribution @xcite and quantum logic gate applications @xcite .
ultrafast lasers are a powerful tool not only for machining @xcite but also for the subtle optical modification of materials .
in particular , the direct - write femtosecond laser technique for creating optical waveguides in dielectric media @xcite is an alternative waveguide manufacturing technique that allows the production of low - volume complex three - dimensional optical _ circuits _ ( fig . [ schematic](a ) ) .
this process has been applied to a wide range of passive and active media to create integrated devices such as microfluidic sensors @xcite , waveguide - bragg gratings @xcite and miniature lasers @xcite . because there is no lithography step in this procedure it enables a waveguide circuit to be taken rapidly from concept to a completed device . however , as with all waveguide fabrication processes , the devices are subject to manufacturing imperfections and there has been no previous demonstration that the use of the laser - writing technique can produce waveguides that can operate on single photons without deleterious effects on phase , spatial mode or polarization . in this paper
we report the first application of laser written waveguide circuits to photonic quantum technologies and , using single- and multi - photon interference experiments , we show that such circuits perform as well as lithographically fabricated devices and are an ideal platform for scalable quantum information science .
in conventional lithographically fabricated integrated optical devices , light is guided in waveguides consisting of a core and a slightly lower refractive - index cladding or buffer layers ( in a manner analogous to an optical fibre ) . in the commonly used flame hydrolysis deposition ( fhd )
fabrication method these structures are lithographically described on top of a semiconductor wafer @xcite . by careful choice of core and cladding dimensions and refractive index difference it is possible to design such waveguides to support only a single transverse mode for a given wavelength range .
coupling between waveguides , to realize beamsplitter - like operation , can be achieved when two waveguides are brought sufficiently close together that the evanescent fields overlap ; this architecture is known as the directional coupler . by carefully selecting the separation between the waveguides and the length of the interaction region the amount of light coupled from one waveguide into the other ( the coupling ratio @xmath0 , where @xmath1 is equivalent to beamsplitter reflectivity ) and its dependance on wavelength can be tuned .
a similar approach can be taken in the case of directly written waveguides , where the waveguide core is formed by local modification of silica @xcite ( or other materials ) . however , unlike the lithographic approach , direct - write circuits can be straightforwardly written in 3d .
we fabricated two chips with a number of direct - write quantum circuits ( dwqcs ) composed of 2@xmath22 directional couplers ( fig .
[ schematic](c ) ) and mach - zehnder interferometers .
the circuits were written inside high purity fused silica using a tightly focused 1 khz repetition rate , 800 nm , 120 fs laser and motion control system similar to that previously reported @xcite .
the writing laser beam was circularly polarised and passed through a 520 @xmath3 m slit before being focused 170 @xmath3 m below the surface of the glass using a 40@xmath2 0.6 numerical aperture microscope objective that was corrected for spherical aberrations at this depth ( fig .
[ schematic](a ) ) .
the writing process created approximately gaussian profile waveguides which were characterised using a near - field refractive index profilometer ( from rinck elektronik ) and a magnifying beam profiler .
the _ measured _ refractive index profile of a typical waveguide is displayed in fig .
[ schematic](b ) and shows a peak refractive index change of 4.55@xmath4 , an average @xmath5 width of 5.5 @xmath3 m and a @xmath6 width ratio of 0.93 . at their design wavelength of 806 nm
these waveguides supported a single transverse mode with orthogonal ( intensity ) @xmath5 widths of 6.1 @xmath3 m @xmath2 6.1 @xmath3 m ( the small amount of measured ri shape - asymmetry not being evident in the guided mode ) .
the design of the directional couplers was functionally identical except for the length of the central interaction region that was varied from 400 to 2000 @xmath3 m to achieve different coupling ratios .
the curved regions of the waveguides were of raised - sine form and connected the input and output waveguide pitch of 250 @xmath3 m down to the closely spaced evanescent coupling region of the devices .
the mach - zehnder interferometers design comprised two 50:50 directional couplers separated by identical 1500 @xmath3 m long arms .
the purpose of these devices was to test the stability of both the completed waveguide circuits and the laser writing system ( which was required to remain stable for the several hours required for fabrication of the directional couplers and interferometers ) .
( a ) -axis ri profile to the right of the peak is an artifact of the measurement method .
( c ) a schematic of an array of directional couplers fabricated by _
fs _ direct writing in a single fused silica chip and an optical micro - graph showing the central coupling region where the waveguides are separated by 10 @xmath3 m .
( d ) measurement setup based on spontaneous parametric down conversion of a continuous wave 402 nm laser diode .
, scaledwidth=100.0% ] ( b ) -axis ri profile to the right of the peak is an artifact of the measurement method .
( c ) a schematic of an array of directional couplers fabricated by _
fs _ direct writing in a single fused silica chip and an optical micro - graph showing the central coupling region where the waveguides are separated by 10 @xmath3 m .
( d ) measurement setup based on spontaneous parametric down conversion of a continuous wave 402 nm laser diode .
, scaledwidth=100.0% ] ( c ) -axis ri profile to the right of the peak is an artifact of the measurement method .
( c ) a schematic of an array of directional couplers fabricated by _
fs _ direct writing in a single fused silica chip and an optical micro - graph showing the central coupling region where the waveguides are separated by 10 @xmath3 m .
( d ) measurement setup based on spontaneous parametric down conversion of a continuous wave 402 nm laser diode .
, scaledwidth=100.0% ] ( d ) -axis ri profile to the right of the peak is an artifact of the measurement method .
( c ) a schematic of an array of directional couplers fabricated by _
fs _ direct writing in a single fused silica chip and an optical micro - graph showing the central coupling region where the waveguides are separated by 10 @xmath3 m .
( d ) measurement setup based on spontaneous parametric down conversion of a continuous wave 402 nm laser diode .
scaledwidth=100.0% ] the quantum circuits were studied using photons obtained using spontaneous parametric down conversion ( spdc ) sources .
for the two - photon interference work the output from a continuous wave ( cw ) 402 nm laser diode was down converted into two unentangled 804 nm photons in a type - i phase matched bibo crystal ( fig .
[ schematic](d ) ) .
the photon pairs were passed through 2 nm bandpass interference filters to improve indistinguishability before being coupled into two polarization maintaining single mode optical fibres ( pmfs ) .
the path length difference could be varied with the @xmath3 m actuator which adjusted the arrival time of the photons with respect to each other at the waveguide chip .
photons were collected from the chip in single mode fibres ( smfs ) and coupled to avalanche photo diodes ( apds ) which were in turn connected to a photon counting and coincidence logic circuit .
index matching oil was used between the fibres and the device under test to reduce fresnel reflections that contribute to coupling losses .
the circuit devices had typical transmission efficiencies of 50% ( which includes coupling and propagation losses ) . in the case of an ideal @xmath7 directional coupler , if two non - degenerate ( _ i.e. _ entirely distinguishable ) photons are coupled into waveguides @xmath8 and @xmath9 ( fig .
[ schematic](d ) ) the photons have a 50% probability of both being transmitted or reflected inside the coupler_i.e .
_ , they behave like classical " particles such that there is a 50% probability of detecting them simultaneously at the two different apds . in contrast , if two degenerate photons are input to @xmath8 and @xmath9 the pair will be transformed into an entangled superposition of two photons in output waveguide @xmath10 and two photons in output waveguide @xmath11 : @xmath12 ) . in our setup , the relative arrival time of the two photons was the free parameter in photon distinguishability and allowed the level of quantum interference to be directly controlled .
more generally the visibility of the hom dip is also limited by other degrees of freedom including polarization , frequency , spatial mode , and beam coupling ratio ( or reflectivity ) of the directional coupler .
while the reflectivity parameter is inherent to the optical circuitry , the remaining degrees of freedom can be attributed to both the source of photons and individual manipulation of these parameters within the circuit . considering the device separately to the source of photons , the _ ideal _ hom dip visibility ,
@xmath13 , is a function of the equivalent beamsplitter reflectivity @xmath1 : @xmath14 which is a maximum for @xmath7 .
imperfections in the waveguides that perturb the state of a photon in any degree of freedom will degrade the degeneracy of the photon pairs and reduce the measured @xmath15 below @xmath16 . assuming a spdc source prepares identical photon pairs the relative visibility @xmath17 of the hom dip provides the operational fidelity of a directional coupler and thus @xmath18 is the key quantifier of the performance of a photonic quantum circuit in preserving photon degeneracy . in this work
we have not corrected for imperfections in either the sources degeneracies or the measurement setups , hence our measurements of @xmath18 are lower bounds for the fidelities of the devices .
in addition to high - fidelity quantum interference ( as quantified by @xmath18 ) , quantum circuits require high - visibility , stable classical interference . using the 2-photon spdc source as a convenient source of 804 nm photons we measured the effective reflectivity of the mach - zehnder interferometers by blocking one arm of the spdc source to input one photon at a time into the arms of the interferometer . in the case of an ideal null - phase difference interferometer with 50:50 beam splitters
the effective reflectivity of the device , @xmath19 , should be 1 .
values for @xmath20 or instabilities under changing environmental conditions can therefore be used as a device performance metric .
quantum logic gates @xcite , quantum metrology @xcite , photon number filters @xcite , entanglement filters @xcite , and biphoton qutrit unitaries @xcite .
when two photons are input into @xmath8 and one in @xmath9 an ideal @xmath21 reflectivity coupler will generate the three - photon entangled state : @xmath22 term .
an analogue of the hom dip can therefore be observed by measuring the rate of detecting two photons in c and one in d as a function of the delay time between the photon in @xmath9 and the two photons in @xmath8 @xcite . to observe a @xmath23 hom dip , as described by eq .
( [ 21hom ] ) , we used a pulsed laser system ( fig . [ 21dip](a ) ) to generate four photons in 2 modes at 780 nm where the dwqcs are also single moded .
the output of a @xmath24150 fs , 80 mhz repetition rate 780 nm ti : sapphire laser was frequency doubled to 390 nm and then down converted into pairs of 780 nm photons in a type - i phase matched bibo crystal .
the photon pairs passed through 3 nm bandpass interference filters before being coupled into two polarization maintaining single mode optical fibres . in all other
respects the setup is similar to the cw one shown in fig .
[ schematic](d ) . by using a fused pmf splitter and single photon apd
we were able to probabilistically prepare the @xmath25 state at input @xmath9 ( fig .
[ 21dip](a ) ) . using a smf fibre coupler
we were able to probabilistically detect two photons at output @xmath10 .
the low probability of preparing the required 4-photon state at the source and the @xmath26 success rate of experimental setup significantly reduced the measurement count rate from that of the 2-photon experiment . by testing circuit stability over long durations ,
interference measurements of this form are a crucial trial of these circuits for applications in advanced , multi - photon quantum optics .
each set of measurements with this setup took @xmath2460 hours to complete and are the first demonstration of interference between more than 2-photons in an integrated waveguide - chip platform .
we measured the performance of the dwqc devices using the 2-photon setup shown schematically in fig .
[ schematic](d ) .
figure [ dip ] shows the raw data for a hom dip in a coupler with @xmath27 ( maximum theoretical visibility @xmath28 ) . the measured visibility is @xmath29 .
figure [ vis ] shows the measured visibility @xmath15 as a function of the equivalent reflectivity @xmath1 for eight couplers on the two chips . the curve is a fit of eq .
( [ visibility ] ) , modified to include a single parameter to account for mode mismatch @xcite .
the average relative visibility for these eight couplers is @xmath30 , demonstrating high performance across all couplers on both chips . .
error bars are determined from fits such as those in fig .
[ dip ] and are comparable to the point size.,scaledwidth=45.0% ] using single - photons , we measured the effective reflectivity of a mach - zehnder interferometer and found that the reflectivity of the device was @xmath31 .
this indicates that the error in written phase shift in the interferometer was very close to zero ( of the order 10 nm ) . using the 4-photon source shown in fig .
[ 21dip](a ) we were able to generate the three - photon entangled state described in eqn .
( [ 21hom ] ) .
figure [ 21dip](b ) shows the generalized hom dip observed in a @xmath32 reflectivity dwqc coupler .
the visibility of this dip is @xmath33 which surpasses the value of @xmath34 previously observed in a bulk optical implementation @xcite ) .
we believe our visibility to be limited by a small amount of temporal distinguishability of the photons produced in the source , nominally in the @xmath35 state , and not by the waveguide device .
these results demonstrate the enhanced stability afforded by guided - wave circuits and are the first demonstration of multi - photon interactions in an integrated optics platform .
( a ) ( b )
dwqcs overcome many of the limitations of standard lithographic approaches : while the devices described here were written in 2d , extension to a 3d architecture is straightforward ; single devices can be made with short turnaround for rapid prototyping ; and the direct - write technique easily produces devices with circular mode profiles , the size and elipticity of which can be adjusted by changing the laser focusing conditions . this ability to tailor the guided mode will enable production of waveguides that better match fibre modes and reduce photon losses , and could be combined with the ongoing development of waveguide @xcite or fibre @xcite based photon sources .
fabrication of sophisticated integrated quantum photonic circuits can now be achieved with only an ultrafast laser system rather than state - of - the - art semiconductor processing facilities .
quantum circuits for photons , regardless of their application , comprise multi - path and multi - photon interferometers involving generalized quantum interference at a beam splitter ( or directional coupler ) .
high visibility interference such as that demonstrated here is therefore crucial to realizing future quantum technologies and the next generation of fundamental quantum optics experiments .
* acknowledgments * + this work was supported by the australian research council through their centres of excellence program , the uk epsrc , qip irc , the macquarie university research innovation fund and the us iarpa . | we report photonic quantum circuits created using an ultrafast laser processing technique that is rapid , requires no lithographic mask and can be used to create three - dimensional networks of waveguide devices .
we have characterized directional couplers the key functional elements of photonic quantum circuits and found that they perform as well as lithographically produced waveguide devices .
we further demonstrate high - performance interferometers and an important multi - photon quantum interference phenomenon for the first time in integrated optics .
this direct - write approach will enable the rapid development of sophisticated quantum optical circuits and their scaling into three - dimensions .
10 url # 1`#1`urlprefix[2][]#2 m. a. nielsen and i. l. chuang , _ quantum computation and quantum information _
( cambridge university press , 2000 ) .
v. giovannetti , s. lloyd , and l. maccone , `` quantum - enhanced measurements : beating the standard quantum limit , '' science * 306 * , 13301336 ( 2004 ) .
a. n. boto , p. kok , d. s. abrams , s. l. braunstein , c. p. williams , and j. p. dowling , `` quantum interferometric optical lithography : exploiting entanglement to beat the diffraction limit , '' phys .
rev .
lett . * 85*(13 ) , 27332736 ( 2000 ) .
n. gisin , g. ribordy , w. tittel , and h. zbinden , `` quantum cryptography , '' rev
. mod .
phys .
* 74 * , 145195 ( 2002 ) .
n. gisin and r. thew , `` quantum communication , '' nat . photon . * 1*(3 ) , 165171 ( 2007 ) . j. l. obrien , `` optical quantum computing , '' science * 318*(5856 ) , 15671570 ( 2007 ) .
a. politi , m. j. cryan , j. g. rarity , s. yu , and j. l. obrien , `` silica - on - silicon waveguide quantum circuits , '' science * 320*(5876 ) , 646649 ( 2008 ) . k. m. davis , k. miura , n. sugimoto , and k. hirao , `` writing waveguides in glass with a femtosecond laser , '' opt . lett . *
21*(21 ) , 17291731 ( 1996 ) .
s. nolte , m. will , j. burghoff , and a. tuennermann , `` femtosecond waveguide writing : a new avenue to three - dimensional integrated optics , '' appl . phys .
a * 77*(1 ) , 109111 ( 2003 ) .
e. knill , r. laflamme , and g. j. milburn , `` a scheme for efficient quantum computation with linear optics , '' nature * 409*(6816 ) , 4652 ( 2001 ) .
a. j. shields , `` semiconductor quantum light sources , '' nat .
photon . * 1*(4 ) , 215223 ( 2007 ) .
h. takesue , s. w. nam , q. zhang , r. h. hadfield , t. honjo , k. tamaki , and y. yamamoto , `` quantum key distribution over a 40-db channel loss using superconducting single - photon detectors , '' nat . photon . * 1*(6 ) , 343348 ( 2007 ) . e. j. gansen , m. a. rowe , m. b. greene , d. rosenberg , t. e. harvey , m. y. su , r. h. hadfield , s. w. nam , and r. p. mirin , `` photon - number - discriminating detection using a quantum - dot , optically gated , field - effect transistor , '' nat . photon . * 1*(10 ) , 585588 ( 2007 ) . g. j. pryde , j. l. obrien , a. g. white , and s. d. bartlett , `` demonstrating superior discrimination of locally prepared states using nonlocal measurements , '' phys .
rev .
lett .
* 94*(22 ) , 220406 ( 2005 ) .
t. yamamoto , k. hayashi , s. k. ozdemir , m. koashi , and n. imoto , `` robust photonic entanglement distribution by state - independent encoding onto decoherence - free subspace , '' nat .
photon . * 2*(8 ) , 488491 ( 2008 ) .
m. w. mitchell , j. s. lundeen , and a. m. steinberg , `` super - resolving phase measurements with a multiphoton entangled state , '' nature * 429*(6988 ) , 161164 ( 2004 ) .
t. nagata , r. okamoto , j. l. obrien , k. sasaki , and s. takeuchi , `` beating the standard quantum limit with four - entangled photons , '' science * 316*(5825 ) , 726729 ( 2007 ) .
k. j. resch , k. l. pregnell , r. prevedel , a. gilchrist , g. j. pryde , j. l. obrien , and a. g. white , `` time - reversal and super - resolving phase measurements , '' phys .
rev .
lett . * 98*(22 ) , 223601 ( 2007 ) .
m. dangelo , m. v. chekhova , and y. shih , `` two - photon diffraction and quantum lithography , '' phys .
rev .
lett . * 87*(1 ) , 013602 ( 2001 ) .
a. s. clark , j. fulconis , j. g. rarity , w. j. wadsworth , and j. l. obrien , `` all - optical - fiber polarization - based quantum logic gate , '' phys .
rev .
a * 79*(3 ) , 030303 ( 2009 ) .
r. r. gattass and e. mazur , `` femtosecond laser micromachining in transparent materials , '' nat .
photon . * 2*(4 ) , 219225 ( 2008 ) .
r. osellame , v. maselli , r. m. vazquez , r. ramponi , and g. cerullo , `` integration of optical waveguides and microfluidic channels both fabricated by femtosecond laser irradiation , '' appl .
. lett . *
90*(23 ) , 231118 ( 2007 ) .
g. d. marshall , m. ams , and m. j. withford , `` direct laser written waveguide - bragg gratings in bulk fused silica , '' opt .
lett . * 31*(18 ) , 26902691 ( 2006 ) .
g. d. marshall , p. dekker , m. ams , j. a. piper , and m. j. withford , `` directly written monolithic waveguide laser incorporating a distributed feedback waveguide - bragg grating , '' opt .
lett .
* 33*(9 ) , 956958 ( 2008 ) .
j. w. chan , t. r. huser , s. h. risbud , and d. m. krol , `` modification of the fused silica glass network associated with waveguide fabrication using femtosecond laser pulses , '' appl .
phys .
a * 76*(3 ) , 367372 ( 2003 ) .
d. little , m. ams , p. dekker , g. marshall , j. dawes , and m. withford , `` femtosecond laser modification of fused silica : the effect of writing polarization on si - o ring structure , '' opt .
express * 16*(24 ) , 2002920037 ( 2008 ) .
m. ams , g. d. marshall , d. j. spence , and m. j. withford , `` slit beam shaping method for femtosecond laser direct - write fabrication of symmetric waveguides in bulk glasses , '' opt .
express * 13*(15 ) , 56765681 ( 2005 ) .
m. ams , g. d. marshall , and m. j. withford , `` study of the influence of femtosecond laser polarisation on direct writing of waveguides , '' opt . express * 14*(26 ) , 1315813163 ( 2006 ) . c. k. hong , z. y. ou , and l. mandel , `` measurement of subpicosecond time intervals between two photons by interference , '' phys .
rev . lett . * 59 * , 20442046 ( 1987 ) .
k. sanaka , t. jennewein , j. pan , k. resch , and a. zeilinger , `` experimental nonlinear sign shift for linear optics quantum computation , '' phys .
rev .
lett . * 92*(1 ) , 017902 ( 2004 ) .
b. h. liu , f. w. sun , y. x. gong , y. f. huang , z. y. ou , and g. c. guo , `` demonstration of the three - photon de broglie wavelength by projection measurement , '' phys . rev .
a * 77*(2 ) , 023815 ( 2008 ) .
k. sanaka , k. j. resch , and a. zeilinger , `` filtering out photonic fock states , '' phys .
rev .
lett .
* 96*(8 ) , 083601 ( 2006 ) .
k. j. resch , j. l. obrien , t. j. weinhold , k. sanaka , b. p. lanyon , n. k. langford , and a. g. white , `` entanglement generation by fock - state filtration , '' phys .
rev .
lett . * 98*(20 ) , 203602 ( 2007 ) . h. f. hofmann and s. takeuchi , `` quantum filter for nonlocal polarization properties of photonic qubits , '' phys .
rev .
lett . * 88*(14 ) , 147901 ( 2002 ) .
r. okamoto , j. l. obrien , h. f. hofmann , t. nagata , k. sasaki , and s. takeuchi , `` an entanglement filter , '' science * 323*(5913 ) , 483485 ( 2009 ) .
b. p. lanyon , t. j. weinhold , n. k. langford , j. l. obrien , k. j. resch , a. gilchrist , and a. g. white , `` manipulating biphotonic qutrits , '' phys .
rev .
lett . * 100*(6 ) , 060504 ( 2008 ) .
p. p. rohde , g. j. pryde , j. l. obrien , and t. c. ralph , `` quantum - gate characterization in an extended hilbert space , '' phys .
rev .
a * 72*(3 ) , 032306 ( 2005 ) .
g. fujii , n. namekata , m. motoya , s. kurimura , and s. inoue , `` bright narrowband source of photon pairs at optical telecommunication wavelengths using a type - ii periodically poled lithium niobate waveguide , '' opt .
express * 15*(20 ) , 1276912776 ( 2007 ) .
q. zhang , x. p. xie , h. takesue , s. w. nam , c. langrock , m. m. fejer , and y. yamamoto , `` correlated photon - pair generation in reverse - proton - exchange ppln waveguides with integrated mode demultiplexer at 10 ghz clock , '' opt . express * 15*(16 ) , 1028810293 ( 2007 ) .
j. fulconis , o. alibart , j. l. obrien , w. j. wadsworth , and j. g. rarity , `` nonclassical interference and entanglement generation using a photonic crystal fiber pair photon source , '' phys .
rev .
lett . * 99*(12 ) , 120501 ( 2007 ) . |
quantum transport property of two - dimensional(2d ) systems in strong magnetic fields has been one of the central issues of the condensed matter physics .
much work has been performed in connection with the quantum hall effect @xcite .
recent discovery of the quantum hall effect in graphene @xcite also stimulates the theoretical interest .
when the system is in a strong magnetic field , the so - called edge states are formed along the boundaries of the system @xcite .
the edge state corresponds to the classical skipping orbit along the boundary and is known to be less influenced by impurities and defects of the system @xcite .
it thus plays an important role in the quantum transport of a two - dimensional system in a strong magnetic field .
numerical studies of the electron states of two - dimensional systems in a strong magnetic field in the presence of boundaries have already been performed by many authors @xcite .
it has been shown that the edge state is well defined and is extended along the boundary as long as its energy lies away from those of the bulk landau subbands and the magnetic fields are strong enough .
on the other hand , when its energy lies in the middle of the bulk landau subbands , the edge state mixes with the bulk states by the impurity scattering , which leads to the localization of edge states @xcite .
recently , it has been demonstrated for the case of long quantum hall wires with uncorrelated disorder potential that the conductance indeed vanishes when the fermi energy is close to the centers of the bulk landau subbands @xcite .
the vanishing conductance can be understood as the consequence of the mixing between the bulk states and the edge states having opposite current directions at each end of the system , as has been confirmed by a quantitative comparison with analytical results .
this transition between a quantized conductance and the insulator is called the chiral metal - insulator transition(cmit ) @xcite . in the present paper ,
the effect of potential correlation in such systems is investigated numerically .
the potential correlation is , in general , important for the transport in low dimensions , since in some cases it changes the localization behavior drastically . for a carbon nanotube
, it has been shown that the potential correlation leads to the absence of back scattering @xcite . even for pure 1d systems , dimer type correlation or long - range correlation of potential
gives rise to delocalized states @xcite .
for the two - dimensional bulk system in a magnetic field , the effect of potential correlation has also been studied based on the continuum model @xcite as well as the tight - binding model @xcite .
it has been demonstrated that the mixing between edge states is suppressed by the potential correlation @xcite . in the analysis on the critical states of the landau bands
, it is seen that the critical energies are insensitive to the potential correlation as long as the strength of the disorder is weak @xcite .
it may then be expected naively that cmit observed for the uncorrelated potential is suppressed and the quantized conductance steps are recovered when the potential correlation is introduced .
we find indeed the suppression of cmit which yields the recovery of the quantized conductance steps .
quite unexpectedly , however , it is found that the localization length associated with the edge states is suppressed by the potential correlation .
this causes the shift of the quantized conductance steps toward higher energies .
it is remarkable that the potential correlation suppresses the conduction of wires , since the potential correlation is usually expected to enhance the conduction of the system .
we find that the scale of the shift is of the order of the strength of random potential and is insensitive to the strength of the magnetic field as long as the landau levels are well defined . on the other hand , no shift appears in the square geometry .
it is therefore necessary to analyze the effect of potential correlation on the edge states in long wires to understand these shifts of conductance plateaus .
we adopt the tight - binding model described by the following hamiltonian on the square lattice @xmath0 where @xmath1 denotes the creation(annihilation ) operator of an electron on the site @xmath2 .
the summation of the phases @xmath3 around a plaquette is equal to @xmath4 , where @xmath5 is the magnetic flux through the plaquette and @xmath6 stands for the flux quantum . the elementary charge and the planck constant are denoted by @xmath7 and @xmath8 , respectively .
all length - scales are measured in units of the lattice spacing .
the site energies @xmath9 are assumed to be distributed with the gaussian probability density @xmath10 and to have the spatial correlation as @xmath11 here @xmath12 and @xmath13 are the parameters which determine the strength of the random potential and the strength of its spatial correlation , respectively . the position vector for the site @xmath2
is denoted by @xmath14 .
the spatially correlated potential is made from the uncorrelated potential @xmath15 s as @xcite @xmath16 when the uncorrelated potential @xmath15 s obey the gaussian distribution with variance @xmath12 , it is easy to verify that @xmath17 satisfies the relations in eqs.([gauss ] ) and ( [ corr1 ] ) .
the parameter @xmath13 , therefore , represents the range of the impurity potential . in the following , we specify the disorder strength by a parameter @xmath18 , since it can be regarded as the effective width of the potential distribution .
we consider a system with the length @xmath19 and the width @xmath20 .
two leads are attached to both ends of the system and the fixed boundary condition is assumed in the transverse direction . for the realization of the isotropic correlation for @xmath17 in the sample region @xmath21 , we consider the additional regions of the width @xmath22 outside of the sample region in performing the summation over @xmath15 in eq.([sum ] ) .
the two - terminal conductance @xmath23 is obtained by means of the landauer formula @xmath24 where @xmath25 is the transmission matrix .
we adopt the transfer matrix method@xcite to evaluate the transmission matrix . throughout the present analysis
, we consider an independent impurity configuration for each value of energy @xmath26 and @xmath13 .
the width @xmath20 is set to be 20 .
the smallest magnetic flux @xmath27 per plaquette is @xmath28 , and the corresponding magnetic length @xmath29 is about @xmath30 , much smaller than the system width @xmath20 .
this leads to the existence of edge states along the boundaries .
we first show the results of conductance for @xmath31 and @xmath32 in fig .
the length of the system is @xmath33 . in the absence of potential correlation ( @xmath34 )
, we see the chiral metal - insulator transition as a function of energies .
the conductance peaks at energies @xmath35 and @xmath36 are the contributions from the lowest and the 2nd lowest landau bands , respectively . it is clearly seen that these peaks disappear for @xmath37 and that the quantized conductance steps are recovered for @xmath38 . it is to be noted here that the recovery starts at @xmath39 . since the contribution from the lowest landau band starts at @xmath40 for the case of @xmath34 , the conductance steps can be understood as being shifted toward higher energies with an amount of @xmath41 in the presence of long potential correlation . , @xmath42 and @xmath33 as a function of the range of the impurity potential @xmath13 and the fermi energy @xmath43 . in the present system ,
ratio of the with @xmath20 to the magnetic length @xmath44 is @xmath45 .
[ fig1 ] ] in order to see the shift more clearly , we examine the conductance in the cases of stronger magnetic fields @xmath46 and @xmath47 . in fig .
[ fig2 ] , the results for @xmath46 are shown .
the length of systems is assumed to be @xmath48 .
here it is clearly seen that the conductance plateaus shift toward higher energies as the correlation length @xmath13 is increased . in both cases ,
the scale of the shifts in the presence of the long correlation turns out to be approximately @xmath49 insensitive to the magnetic fields .
we also evaluate the conductance for @xmath50 , namely for the square system .
it is then found that no shift appears in that case .
this is consistent with the result for the 2d bulk system that the levitation of critical states due to the potential correlation for the present range of disorder strength ( @xmath51 ) is much smaller than the present scale of the shifts of conductance plateaus @xcite .
the present shift of conductance plateaus is thus a specific feature to long wires , suggesting that the shift arises from the one - dimensional character of the system rather than the 2d bulk properties .
apart from the shift of plateaus , it is also confirmed that cmit fades away for larger values of @xmath52 .
detailed energy dependencies of the conductance in the presence of potential correlation ( @xmath53 ) and that in its absence ( @xmath54 ) are shown in fig . [ fig3 ] .
it is clearly seen that the shifts are common to these three plateaus and cmit observed for @xmath54 @xcite disappears for @xmath55 .
the random jumps between the quantized plateaus for @xmath55 are fluctuations due to different realizations of disorder potential .
it should be emphasized that the conductance plateau transitions occur away from the peaks of the density of states in the case of @xmath56 . for @xmath57 and @xmath58 as a function of the potential range @xmath13 and the fermi energy @xmath43 .
the magnetic flux @xmath27 is assumed to be @xmath59 .
the magnetic length @xmath60 and accordingly @xmath61 .
[ fig2 ] ] for @xmath54 ( @xmath62 ) and @xmath63 ( @xmath64 ) in the case of @xmath57 , @xmath65 and @xmath58 . for @xmath56 , @xmath66
is plotted instead of @xmath67 .
conductance @xmath68 for @xmath69 is also calculated and @xmath70 is plotted as a solid line .
the density of states ( dos ) for @xmath56 obtained by the green s function method @xcite is plotted as a dotted curve .
[ fig3 ] ] in order to clarify how the scale of the shift depends on @xmath71 , we calculate the conductance also for @xmath72 and @xmath73 in the case of @xmath58 and @xmath74 .
we obtain the shift to be @xmath75 and @xmath76 , respectively .
these results suggest that the scale of the shift is in proportion to @xmath71 .
the shift of plateaus means that the edge states which are transmitted in the case of @xmath34 are reflected for @xmath77 .
the localization length in the case of @xmath77 must therefore be much smaller than that in the case of @xmath34 .
this is surprising since the potential correlation normally reduces the electron localization .
we thus examine the effect of potential correlation on the localization length along the wire in the presence of edge states .
the localization length @xmath78 estimated by the transfer matrix method @xcite is shown in fig .
[ fig4 ] for @xmath46 and various values of @xmath13 . here
we clearly see that the localization length @xmath78 is much shorter than the system length @xmath79 for the energies @xmath80 in the presence of potential correlation .
this is consistent with the vanishing conductivity in the presence of long potential correlation for energies lower than @xmath81 in figs .
[ fig2 ] and [ fig3 ] . , @xmath82 , @xmath83 and @xmath84 .
the disorder and the magnetic field are assumed to be @xmath57 and @xmath85 , respectively .
the magnetic length is @xmath60 .
now we show the present shift of conductance plateaus can be understood as a semi - classical effect .
it is useful to recall here the fact that the edge state is reflected by the potential barrier when its energy measured from the corresponding landau level @xmath86 is smaller than the energy of the potential barrier @xcite , where @xmath87 denotes the cyclotron frequency .
this suggests that the shift of conductance plateaus also occurs in the presence of potential barrier across the wire instead of disorder potential .
we have confirmed in the present lattice model that the shifts expected in the continuum model indeed occur when the thickness of the barrier is larger than the magnetic length . with this property of the edge states ,
it is natural to expect that the edge states having lower energies ( @xmath88 ) are deformed considerably by the correlated potential varying with scales larger than the magnetic length , because such a potential would act as a local potential barrier . due to the deformation of the edge state trajectory , in certain region of a long sample ,
the effective width of the sample becomes narrow enough to induce the mixing of edge states with opposite directions , leading to the reflection .
as the potential correlation is increased , the deformation would become larger and the probability for the reflection of edge states is likely to increase , accordingly .
the same order of the shift ( @xmath89 ) in each plateau transition is also naturally understood since @xmath90 determines the probability of reflection .
the above semi - classical argument is consistent with our results that the shift becomes significant when the potential range @xmath91 is large enough and that in the long correlation limit , it saturates around @xmath92 which is effectively the maximum of the potential energy ( fig .
[ fig2 ] ) .
it would also account for the fact that the contributions from the lowest and the 2nd lowest landau levels vanish when the potential range @xmath13 approaches to the order of the magnetic length in fig .
[ fig1 ] . in the presence of long potential correlation
, it is thus expected that the conductance steps occur at critical energies @xmath93 in long wires . for certain fixed impurity configurations ,
we have observed complex structure around the critical energy , which is expected to be the resonant tunneling @xcite between edge states .
it is important to note that the mixing of the edge states at one place of the wire would be enough to reflect the whole current associated with them .
the probability to have such a place in a particular sample apparently depends on the length of the wire .
it is natural that there is no shift in the case of the short system ( @xmath50 ) since such probability is very small .
the probability @xmath94 to have high potential regions across the wire in a sample is estimated as @xmath95 for a small @xmath96 which is the probability that the potential energy in the box of the size @xmath13 is larger than @xmath90 . requiring @xmath94 is to be of order of unity , we get @xmath97 which suggests that @xmath20 must be practically several times @xmath13 or less to observe these phenomena . in reality , when the magnetic field is @xmath98 , the magnetic length is @xmath99 , and the present systems correspond to wires whose width @xmath20 is in the range @xmath100 .
the present numerical results suggest that the shift of conductance plateaus can be observed when @xmath101 and @xmath102 . in summary
, we have investigated the effect of potential correlation in quantum hall wires .
it has been shown clearly that the potential correlation shifts the quantized conductance plateaus toward higher energies .
the scale of the shifts is of the order of the strength of the random potential .
this shift is specific to systems with the wire geometry and is insensitive to the strength of the magnetic fields .
this phenomenon is related to the transport property of edge states in the presence of long potential correlation .
we have argued that the potential correlation enhances the mixing of the edge states at the opposite edges , which yields the reflection of edge channels . for correlated potential , the positions of the conductance plateau transitions in quantum hall wires do not necessarily coincide with the positions of the bulk landau levels .
the chiral metal - insulator transition is absent when the potential correlation is much larger than the magnetic length of the system .
this work was partly supported by a grant - in - aid no .
16540294 and no .
18540382 from the japan society for the promotion of science and by `` schwerpunktprogramm quanten - hall - systeme '' , grant no .
ke 807/2 - 1 from the german research council ( dfg ) .
10 t. ando , prog .
. suppl . * 84 * , 69 ( 1985 ) .
h. aoki , rep .
. phys . * 50 * , 655 ( 1987 ) .
edited by r.e .
prange and s.m .
girvin , ( springer - verlag , 1987 ) , 2nd ed .
y. hatsugai , j. phys . : condens .
matter * 9 * , 2507 ( 1997 ) .
b. huckestein , rev .
67 * , 357 ( 1995 ) .
b. kramer , t. ohtsuki , and s. kettemann , phys .
rep . * 417 * , 211 ( 2005 ) .
novoselov et al . ,
nature * 438 * , 197 ( 2005 ) ; y. zhang et al .
, nature * 438 * , 201 ( 2005 ) .
halperin , phys .
rev . * b25 * , 2185 ( 1982 ) .
m. bttiker , phys . rev . *
b38 * , 9375 ( 1988 ) .
t. martin and s. feng , phys .
lett . * 64 * , 1971 ( 1990 ) ; phys . rev . *
b44 * , 9084 ( 1991 ) .
l. schweitzer , b. kramer and a. mackinnon , j. phys . * c17 * , 4111 ( 1984 ) .
t. ohtsuki and y.ono , solid state commun .
* 65 * , 403 ( 1988 ) ; * 68 * , 787 ( 1988 ) ; j. phys .
* 58 * , 956 ( 1989 ) .
y. ono , t. ohtsuki and b. kramer , j. phys .
. jpn . * 58 * , 1705 ( 1989 ) .
t. ando , phys . rev .
* b42 * , 5626 ( 1990 ) .
t. ando and h. akera , in _ high magnetic fields in semiconductor physics iii _ , edited by g. landwehr ( springer , heidelberg , 1992 ) , p.291 .
s. kettemann , b. kramer , and t. ohtsuki , jetp lett .
* 80 * , 285 ( 2004 ) .
a. struck , b. kramer , t. ohtsuki , and s. kettemann , phys .
rev . * b72 * , 035339 ( 2005 ) . t. ando and t. nakanishi , j. phys .
. jpn . * 67 * , 1704 ( 1998 ) .
dunlap , h - l .
wu and p.w .
phillips , phys .
lett . * 65 * , 88 ( 1990 ) .
f.a.b.f . de moura and m.l .
lyra , phys . rev .
lett . * 81 * , 3735 ( 1998 ) .
t. ando and y. uemura , j. phys .
soc . jpn . *
36 * , 959 ( 1974 ) .
t. ohtsuki and y. ono , j. phys .
jpn . * 58 * , 3863 ( 1989 ) .
t. ando , j. phys .
. jpn . * 53 * , 3101 ( 1984 ) . th .
koschny and l. schweitzer , phys . rev .
* b67 * , 195307 ( 2003 ) .
pendry , a. mackinnon and p.j .
roberts , proc .
* a437 * , 67 ( 1992 ) .
a. mackinnon and b. kramer , phys .
lett . * 47 * , 1546 ( 1981 ) ; z. phys . *
b53 * , 1 ( 1983 ) .
t. ohtsuki and y. ono , j. phys .
jpn . * 59 * , 637 ( 1990 ) .
y. takagaki , phys . rev .
b 57 , 4009 ( 1998 ) .
j.k . jain and s.a .
kivelson , phys .
lett . * 60 * , 1542 ( 1988 ) .
raikh and t.v .
shahbazyan , phys . rev . *
b51 * , 9682 ( 1995 ) . | quantum transport properties in quantum hall wires in the presence of spatially correlated random potential are investigated numerically .
it is found that the potential correlation reduces the localization length associated with the edge state , in contrast to the naive expectation that the potential correlation increases it .
the effect appears as the sizable shift of quantized conductance plateaus in long wires , where the plateau transitions occur at energies much higher than the landau band centers .
the scale of the shift is of the order of the strength of the random potential and is insensitive to the strength of magnetic fields .
experimental implications are also discussed . |
semiclassical gravity has played a central role in theoretical physics .
phenomena such as the hawking effect or cosmological particle production are commonly considered to be crucial first steps on the way to building up a consistent fully - quantum theory of gravity ( see for example @xcite ) .
however a fundamental limit to these investigations is imposed by the fact that their most basic description is based on linear qft on a fixed ( classical ) continuum spacetime .
several theoretical approaches have been developed to overcome this limitation : in a fashion that we can call `` top - down '' , string models [ brane models ] have in some special situations developed a high energy description of the hawking effect @xcite , while `` bottom - up '' approaches , based on stochastic gravity and the einstein langevin analysis of particle creation by a gravitational field , have in recent years provided further insight @xcite .
on the other hand , the physics community has so far lacked any possibility for direct experimental tests of these ideas .
indeed this lack of experimental guidance is a severe hindrance with respect to further developments in semiclassical gravity ( or full - fledged quantum gravity for that matter ) .
in particular we have no _ experimental _ guidance regarding the manner in which the predictions of curved spacetime quantum field theory are changed once the hypotheses of non - discreteness and/or a non - fluctuating background are relaxed . in this regard
the analogue models of gravity developed in recent years can be considered a first attempt to create an arena which can serve as a theoretical , and possibly observational , laboratory to test aspects of these scenarios .
no experimental set up has yet been realized in which the predictions of analogue models can be observationally tested .
nevertheless theoretical analyses of analogue models @xcite have been so far remarkably successful in teaching us how semiclassical gravity phenomena are sensitive to possible quantum gravity effects , such as , for example , modified ( lorentz violating ) dispersion relations @xcite .
( see for example the trans - planckian problem in the hawking effect @xcite and in cosmology @xcite . )
what we intend to discuss in this article is a particular class of experiment that we hope could be realized in the very near future wherein certain analogue gravity model predictions could be tested .
the interest in doing so would not just be that of confirming a now well - established theoretical prediction , but mainly trying to evince deviations from the naive theoretical predictions due to the intrinsic discrete nature of the experimental system and/or to the possible role of non - linearities .
we shall focus our attention on the analogue gravity system established by the propagation of linearized phase perturbations in a bose - einstein condensate @xcite .
in particular we shall consider an experiment where a time - varying scattering length is used to simulate the cosmological expansion of the universe , and its associated quantum creation of particles .
it is interesting to note that in reference @xcite the authors proposed an explanation of the so called `` bosenova '' phenomenon @xcite ( a controlled instability of the bulk condensate induced by a sign variation of the scattering length ) through a particular implementation of a version of analog cosmological particle production . in that approach the entire bulk of the condensate
is rendered unstable and suffers catastrophic breakdown .
our current paper takes a complementary approach : instead of trying to explain an observed phenomenon via analog cosmological particle production , we shall instead consider the most favourable conditions to observe it preferably without violent disruption of the entire condensate
. the scheme of the paper will be as follows : in the next section we will review the physics of becs regarding its analogue gravity features .
section iii will be devoted to the discussion of how to simulate a frw universe within a bec .
there exist two main routes to do this .
one considers an explosive expansion of the condensate ; the other makes use of the possibility of tuning the strength of the interaction between the different bosons in the condensate .
this latter route will be the main concern of this paper . in section iv we
will first qualitatively describe how the modification of the interaction strength ( encoded in the value of the scattering length ) yields cosmological particle creation .
next , in subsection a we will discuss [ in the context of current bec technology ] whether there exists a regime in which this particle creation process can actually be reproduced .
we will see that there is a limit on the rapidity of change of the background configuration , associated with the need to enforce a `` markov approximation '' , in order for the whole construction to make sense .
however , this bound still leaves a lot of parameter space available to look for the particle creation effect .
subsection b reviews the cosmological particle creation process , emphasizing the particular features of bec systems .
then , subsection c discuss the actual observability of the effect .
finally , we conclude with a short summary and discussion .
bose - einstein condensates ( bec ) have recently become subject of extensive study as possible analogue models of general relativity @xcite .
in particular it was shown that the propagation of phase perturbations in a bec can under certain conditions closely mimic the dynamics of quantum fields in curved spacetimes . in previous papers we noted that a varying scattering length in the bec system corresponds to a varying speed of light in the `` effective metric '' @xcite .
recent experiments have indeed achieved a controlled tuning of the scattering length in rubidium 85 @xcite .
the effect is powerful enough to lead to large non - perturbative changes in the effective metric .
let us start by very briefly reviewing the derivation of the acoustic metric for a bec system . in the dilute gas approximation , one can describe a bose gas through a quantum field @xmath0 satisfying @xmath1 here @xmath2 parameterizes the strength of the interactions between the different bosons in the gas
. it can be re - expressed in terms of the scattering length as @xmath3 as usual , the quantum field can be separated into a macroscopic ( classical ) condensate and a fluctuation : @xmath4 , with @xmath5 .
then , by adopting the self - consistent mean field approximation @xcite @xmath6 one can arrive at the set of coupled equations : @xmath7 @xmath8 here @xmath9 the equation for the classical wave function of the condensate is closed only when the back - reaction effect due to the fluctuations are neglected .
( this back - reaction is hiding in the parameters @xmath10 and @xmath11 . )
this is the approximation contemplated by the gross - pitaevskii equation . in general one will have to solve both equations simultaneously . adopting the madelung representation for the wave function of the condensate @xmath12,\ ] ] and defining an irrotational `` velocity field '' by @xmath13 ,
the gross - pitaevskii equation can be rewritten as a continuity equation plus an euler equation : @xmath14 these equations are completely equivalent to those of an irrotational and inviscid fluid apart from the existence of the so - called quantum potential @xmath15 , which has the dimensions of an energy .
note that @xmath16 = \nabla_j \left [ -{\hbar^2\over4 m } \ ; n_c \ ; \nabla_i \nabla_j \ln n_c \right],\ ] ] which justifies the introduction of the so - called quantum stress tensor @xmath17 this tensor has the dimensions of pressure , and may be viewed as an intrinsically quantum anisotropic pressure contributing to the euler equation .
if we write the mass density of the madelung fluid as @xmath18 , and use the fact that the flow is irrotational then the euler equation takes the form @xmath19 + \rho \ ; \bnabla \left[\frac{v_{ext}(t,\x)}{m } \right ] + \bnabla \left[{\g \rho^2\over 2 m^2}\right ] + \bnabla \cdot \sigma^{\rm quantum } = 0 .
\label{e : euler2}\ ] ] note that the term @xmath20 has the dimensions of specific enthalpy , while @xmath21 represents a bulk pressure .
when the gradients in the density of the condensate are small one can neglect the quantum stress term leading to the standard hydrodynamic approximation . because the flow is irrotational ,
the euler equation is often more conveniently written in hamilton
jacobi form : @xmath22 apart from the wave function of the condensate itself , we also have to account for the [ typically small ] quantum perturbations of the system ( [ quantum - field ] ) .
these quantum perturbations can be described in several different ways , here we are interested in the `` quantum acoustic representation '' @xmath23 where @xmath24 are real quantum fields . by using this representation equation ( [ quantum - field ] ) can be rewritten as @xmath25 here @xmath26 represents a second - order differential operator obtained from linearizing the quantum potential . explicitly : @xmath27\ ; \widehat n_1 + \half n_c^{-1/2 } \;\nabla^2 ( n_c^{-1/2}\ ; \widehat n_1 ) . % \nonumber\\&&\end{aligned}\ ] ] the equations we have just written can be obtained easily by linearizing the gross - pitaevskii equation around a classical solution : @xmath28 , @xmath29 .
it is important to realize that in those equations the back - reaction of the quantum fluctuations on the background solution has been assumed negligible .
we also see in those equations , ( [ pt1 ] ) and ( [ pt2 ] ) , that time variations of @xmath30 and time variations of the scattering length @xmath31 appear to act in very different ways .
whereas the external potential only influences the background equation ( [ e : hj ] ) [ and hence the acoustic metric in the analog description ] , the scattering length directly influences both the perturbation and background equations . from the previous equations for the linearized perturbations it is possible to derive a wave equation for @xmath32 ( or alternatively , for @xmath33 ) .
all we need is to substitute in equation ( [ pt1 ] ) the @xmath33 obtained from equation ( [ pt2 ] ) .
this leads to a pde that is second - order in time derivatives but infinite order in space derivatives to simplify things we can construct the symmetric @xmath34 matrix @xmath35 .
\label{e : explicit}\ ] ] ( greek indices run from @xmath36@xmath37 , while roman indices run from @xmath38@xmath37 . ) then , introducing @xmath39dimensional space - time coordinates @xmath40 the wave equation for @xmath41 is easily rewritten as @xmath42 where the @xmath43 are differential operators acting on space only : @xmath44^{-1 } \\
f^{0j } & = & -\left [ \g(a ) - { \hbar^2\over 2m}\ ; d_2 \right]^{-1}\ ; { \nabla^j \theta_0\over m } \\
f^{i0 } & = & - { \nabla^{i } \theta_0\over m } \ ; \left [ \g(a ) - { \hbar^2\over2 m}\ ; d_2 \right]^{-1 } \\
f^{ij } & = & { n_c \ ; \delta^{ij}\over m } - { \nabla^{i } \theta_0\over m } \ ; \left [ \g(a ) - { \hbar^2\over2 m}\ ; d_2 \right]^{-1}\ ; { \nabla^{j } \theta_0\over m}.\end{aligned}\ ] ] now , if we make an spectral decomposition of the field @xmath45 we can see that for wavelengths larger than @xmath46 ( this corresponds to the `` healing length '' , as we will explain below ) , the terms coming from the linearization of the quantum potential ( the @xmath26 ) can be neglected in the previous expressions , in which case the @xmath43 can be approximated by numbers , instead of differential operators .
( this is the heart of the acoustic approximation . )
then , by identifying @xmath47 the equation for the field @xmath48 becomes that of a ( massless minimally coupled ) quantum scalar field over a curved background @xmath49 with an effective metric of the form @xmath50.\ ] ] here the magnitude @xmath51 represents the speed of the phonons in the medium : @xmath52
to find analog models for cosmological spacetimes we will consider a generalized gp equation where the external potential and the coupling constant can both change with time @xmath53 the technical steps in the calculation change in a straightforward manner and lead to a simple time - dependent acoustic metric @xmath54 \label{fluid - geometry}\end{aligned}\ ] ] it is this time - dependent effective metric that we now wish to use for simulating a cosmological spacetime and , subsequently probing cosmological particle production .
starting from the geometry ( [ fluid - geometry ] ) there are different ways in which one can reproduce a cosmologically expanding geometry . following @xcite one can take a radial profile for the velocity @xmath55 , with @xmath56 a scale factor depending only on @xmath57 , and define a new radial coordinate as @xmath58 . in these new coordinates ,
the metric will be expressed as @xmath59 .
\label{fluid - geometry2}\end{aligned}\ ] ] now , one solution for the bec wave function that reproduces a frw universe is this : introducing a hubble - like parameter , @xmath60 the equation of continuity can be written as @xmath61 with @xmath62constant .
then , the solution for @xmath63 can be obtained from ( [ e : hj ] ) @xmath64 and it requires an external potential of the form @xmath65 one could certainly construct such a potential in a `` sufficiently large '' region around @xmath66 ; this would correspond to a `` sufficiently large '' part of a homogeneous and flat frw universe . outside this region
, the potential will have in practice some confining walls .
the final effective metric can be written as @xmath67 with @xmath68 so finally we end up with a frw universe whose proper friedmann time , @xmath69 , is related to the laboratory time , @xmath57 , by @xmath70 .
although this explosion route seems promising , one should note that this analog models has substantial drawbacks .
in particular it is easy to see from equation ( [ eq : theta - ex ] ) that one would get for the condensate a linearly rising velocity field @xmath71 .
hence this particular realization of a frw effective geometry is guaranteed to possess an apparent horizon , a spherical surface in which the speed of the fluid surpasses the speed of sound . from a dynamical point of view
, this might introduce many practical problems not intrinsically inherent to the type of geometries we are trying to reproduce .
because of this , we view the use of an exploding medium as not being the preferred route for building an analogue for an expanding frw universe .
( for an alternate view , where the explosion route is the primary focus of attention , see @xcite . )
another way to reproduce cosmologically expanding configurations in , we think , a cleaner fashion is by taking advantage of the possibility to change the value of the scattering length offered by some becs @xcite .
let us note that in reference @xcite , the authors already used the existence of time dependence on the scattering length , in combination with a time dependent external potential , to create explosive configuration of the type described in previous subsection . here
, we are going to use this variability with a different strategy .
let us again start from ( [ fluid - geometry ] ) but now with @xmath72 at all times : @xmath73 then , it is not difficult to envisage a situation in which @xmath74 is constant in a `` sufficiently large '' region ( think of a `` sufficiently large '' close - to - hard - walled box ) .
then , the continuity equation is directly satisfied and the hamilton
jacobi equation tells us that with a fixed external potential the phase function @xmath63 will depend only on @xmath57 adapting itself to the changes of @xmath75 . changing
the scattering length with time directly causes changes in the value of the propagation speed @xmath76 .
( that temporal changes in the velocity of sound can be interpreted as a cosmological expansion without invoking any sort of velocity in the medium has already been suggested in @xcite in the context of superfluid helium . )
we now define @xmath77^{1/2 } dt$ ] and write @xmath78 where @xmath79 in this model an expansion corresponds to a decrease of the scattering length and vice versa .
the hubble function for this spacetime is @xmath80 ( the prime represents derivative with respect friedmann time , the dot derivative with respect laboratory time ) .
this is the model we will consider in the following discussion .
let us now present a qualitative explanation of how is that we can closely simulate cosmological particle creation within this model
. we can start with a condensate in a stationary state described , in a sufficiently large volume , by a constant background density @xmath74 and a phase function linear in time , @xmath81 .
this is a solution of the gp equation . for this
, one needs to have a potential that reproduces a large enough close - to - hard - walled box and that satisfies the condition @xmath82 .
apart from this classical background , there will be some quantum fluctuations over it . at temperatures
much below the critical temperature these quantum fluctuations are very small and can be described by the bogoliubov equations .
( these quantum fluctuations are present even at zero temperature owing to the so - called quantum depletion phenomenon , see for example @xcite . )
let us consider that these quantum fluctuations are in their vacuum state .
if one now decreases the value of the scattering length in a sufficiently slow manner ( this issue will be discussed bellow ) , all the individual bosons ( this is only an approximate concept in an interacting theory ) will be affected in the same way .
this means that the value of the background magnitudes these are the coherent magnitudes will be slowly modified .
the gp equation tells you that the density function @xmath74 will be kept fixed while the phase function will develop a non - trivial dependence with time . at the same time
the value of the speed of sound will decrease .
now , apart from the background configuration , what happens with the additional quantum fluctuations ?
the equation satisfied by the quantum fluctuations is , in the acoustic approximation ( that is , for long wavelengths ) that of a massless minimally coupled scalar field over an expanding background and therefore , it will yield cosmological pair production of particles . an interesting point to notice
is that varying the external potential by this we mean changing with time the value of the external potential , @xmath83 , in the central homogeneous region changes the background configuration in the same manner as varying the scattering length does .
however , as the speed of sound does not depend on the external potential , its variations will not lead to cosmological particle creation .
now , what can we say about the observability of the particle production process ?
the standard technique to `` see '' phonons is tomographic imaging .
one opens the trap and looks at the expansion of the condensate .
phonons correspond to distributions for the momentum of the atoms in the trap and different momenta correspond to different travel distances of the atoms after you switch of the trap .
taking snapshots of this evolution shows the density contrasts and then the original momentum distribution .
if the wave function @xmath84 is split into the condensate wave function @xmath85 plus quantum excitations @xmath86 ( in this situation , we mean atoms ) then , the density you observe is @xmath87 as @xmath88 .
therefore , the observability of the effect will depend on the value of the ratio @xmath89 or more simply on the spatially integrated ratio @xmath90 if this quantity is of order unity ( say 1/2 or 1/10 ) then phonons can probably be `` seen '' ; if it is 1/100 then seeing phonons is technologically difficult . if the particle production process was so strong that the calculation of @xmath91 resulted on values close to unity or higher , this would indicate that the bogoliubov mean field approximation would have been violated , and the bec itself disrupted .
let us now perform some explicit calculations of the particle production expected in realistic situations with present - day becs in which the scattering length @xmath31 is changed in time from some initial value to a different final value .
we first have to know how quickly we can drive these temporal changes while still ensuring that the different approximations involved in the analysis remain valid .
the previous analysis shows that in order to consider particle creation driven by a time - varying scattering length we must be sure to work in a regime where the background is `` instantaneously '' reacting to the changes in @xmath31 .
moreover the very derivation of the effective metric description is based on the gp equation which we then want to make sure holds at each instant of time .
so we must first determine the upper bound on the rapidity of the change in the scattering length @xmath31 which still permits the gp equation to hold .
this will also give an upper bound for the frequencies of the quasi - particles that might be created ( if @xmath69 is the shortest timescale over which we can drive the system then @xmath92 is the largest frequency of the quasi - particle we can create ) .
the validity of the gp in describing the bose - einstein condensate is related to the validity of several crucial assumptions which permit us to perform certain approximations on the fundamental multi - particle hamiltonian description .
the relevant approximations are generally stated to be the `` mean - field '' approximation and the dilute gas approximation .
it is nevertheless important to note that in a dynamical situation a third approximation , which we can call `` markovian '' approximation , is implicitly assumed .
let us review the meaning and implications of these approximations : the mean field approximation is based on the assumption that most of the atoms are in the condensate phase and that the influence of the non condensed fraction can be neglected .
this implies that significant creation of quasi - particles with excessive energies can be dangerous .
in particular from the bogoliubov dispersion relation @xcite @xmath93 we can deduce that excitation of quasi - particles with wavelengths comparable to the healing length would led to free particle states ( for @xmath94 ; @xmath95 ; @xmath96 ) .
this argument seems to imply that one should require the typical timescale for the change in @xmath31 to be no shorter than the healing time .
( which is the analogue in this situation of the planck time in quantum gravity . )
the dilute gas approximation is instead related to the way the interaction potential is simplified in the gp equation .
this approximation is valid if @xmath97 , so we shall have to keep the amplitude of the changes in @xmath98 small in order to satisfy this bound .
we wish to emphasise that the dilute gas approximation does not appear to be a crucial approximation for the analog gravity picture to hold .
as long as the interaction term @xmath99 can be generalized to be some higher order ( but still local ) term @xmath100 an analog gravity description is not precluded ( see , for example @xcite ) .
finally the markovian approximation is related to the fact that in dynamical situations the two - body time - dependent scattering matrix can have a complicated form due to the `` memory '' of the system ( see e.g. section iv - a of the paper by khler and burnett @xcite ) .
basically in these situations the system is described by a gp - like equation where the interaction term includes a `` delay term '' described by an integration over different times . the necessary assumption in order to have a markovian description of the dynamics ( which together with the two previous approximations leads to the gp equation ) is then that the timescales on which external parameters are changing are longer that the two - body collisional duration .
that is , longer than the timescale over which a single interaction happens .
( reduced to the bare bones , we are asking that the scattering length does not change significantly during the period when a pair of atoms are interacting . )
we can estimate the two - body collisional time by a simple calculation .
all we need is the typical size of the region of strong interaction of two atoms in the condensate and the typical speed with which they move .
the first quantity can be assumed to be of the order of the van der waals scale length : the inter - atomic potential @xmath101 is characterized by a short - range region of strong chemical bonding and a long - range van der waals potential , @xmath102 this leads to a van der waals scale length @xcite , @xmath103 this length is basically the size of the region of strong interaction : for @xmath104 the scattering wave function oscillates rapidly due to the strong interaction potential . in alkali ground state interactions , @xmath105 is the same for all hyperfine states of a given atomic pair ; consequently , @xmath106 is the same for all collision channels .
for example in the case of na@xmath107 , it is about 2.4 nm . we shall assume here generically @xmath108 nm . regarding the typical speed of the atoms ,
this is set by the de broglie momentum generated by the trap confinement : @xmath109 and @xmath110 .
we shall assume a trap of typical size of ten microns .
we then get @xmath111 we now confront this quantity with the timescale we have to be faster than in order to create modes with wavelengths shorter than the condensate size @xmath112 .
this is @xmath113 .
so @xmath114 for typical bec systems @xmath115m@xmath116 m ( assuming that the scattering length ranges from one to a hundred nanometers ) so @xmath117 note that @xmath118 can be computed to be @xmath119 so a microsecond is the shortest timescale allowed for the change in @xmath31 .
note that this interaction time is shorter than the healing time @xmath120@xmath121 which plays the role of the `` planck time '' in this system .
thus the gp equation is valid in the entire `` sub - healing '' regime , which is the primary regime of interest , and continues to hold well into the `` trans - healing '' regime ( although the previous comments regarding the breakdown of the mean - field approximation in this regime remain valid ) .
now that we have estimated how fast the change in the scattering length can be driven , we can propose a particular time dependence and derive an estimate for the relative production of phonons .
particle production in an expanding universe has been extensively studied in the framework of semiclassical gravity ( see e.g. @xcite ) . in this regard
the scope of this subsection will be to present an example of these calculations to non specialists as well as to evaluate the experimental feasibility of an experimental test . as a test - bed we shall choose a slightly simplified version of parker s model @xcite .
the frw metric with flat spatial sections can be alternatively written as @xmath122 in which we are using for convenience a special type of pseudotime @xmath123 with @xmath124 .
the scale factor is independent of @xmath125 hence the mode decomposition for the quantum scalar field can be written as @xmath126 where the @xmath127 are solutions of the equation @xmath128 that satisfy the normalization condition @xmath129 imagine now that the scale factor @xmath130 undergoes a finite amount of expansion in a monotonic fashion .
this means that the scale factor passes from and initial value @xmath131 at @xmath132 to a final value @xmath133 at @xmath134 ( or what is the same , that the scattering length passes from and initial value @xmath135 to a final value @xmath136 ) .
spacetime can be approximated at early times and at late times respectively by two different minkowski spacetimes . as is standard
, we will assume that the quantum scalar field is initially in the vacuum state associated to the initial minkowski spacetime .
then , we want to calculate the particle content of this state in the final minkowski spacetime .
there is a particularly convenient choice of function @xmath137 for which the physics is clear and for which one can exactly solve equation ( [ mode - equation ] ) : @xmath138 .
\label{eq : prof } \end{aligned}\ ] ] ( this is a slight simplification of the model considered in parker @xcite ; a variation of this model can be found in the book by birrell and davies @xcite , see pp .
if we now impose as boundary condition for a solution that at @xmath139 @xmath140 one finds a particular set of exact solutions @xcite @xmath141 \ ; [ 2\cosh(\eta/\eta_0)]^{-ik\eta_0(a_{sf}^2-a_{si}^2)/2 } \nonumber\\ & & \times \ ; _ 2f_1\left(1+ik\eta_0[a_{sf}^2-a_{si}^2]/2,+ik\eta_0[a_{sf}^2-a_{si}^2]/2 ; 1-ik\eta_0 a_{si}^2 ; { 1\over2}[1+\tanh(\eta/\eta_0 ) ] \right).\end{aligned}\ ] ] these solutions correspond to positive _ in - going _ modes . similarly , the precise form of the exact solutions that at @xmath142 admit the asymptotic form @xmath143 ( the positive _ out - going _ modes ) is @xmath144 \ ; [ 2\cosh(\eta/\eta_0)]^{ik\eta_0(a_{sf}^2-a_{si}^2)/2 } \nonumber\\ & & \times \ ; _ 2f_1\left(1-ik\eta_0[a_{sf}^2-a_{si}^2]/2,-ik\eta_0[a_{sf}^2-a_{si}^2]/2 ; 1-ik\eta_0 a_{si}^2 ; { 1\over2 } [ 1-\tanh(\eta/\eta_0 ) ] \right).\end{aligned}\ ] ] the bogoliubov coefficients relating the early time ( _ in - going _ ) and late time ( _ out - going _ ) bases are then defined as @xmath145 .
\label{bogo}\end{aligned}\ ] ] from this expression , and the exact solutions above , one can show @xcite that @xmath146/2 ) \ ;
\gamma(-ik\eta_0[a_{sf}^2+a_{si}^2]/2 ) } , \\ \beta_k & = & -{2 a_{sf } a_{si}\over a_{sf}^2-a_{si}^2 } \ ; { \gamma(-ik\eta_0 a_{si}^2 ) \ ; \gamma(+ik\eta_0 a_{sf}^2 ) \over
\gamma(+ik\eta_0[a_{sf}^2-a_{si}^2]/2 ) \ ; \gamma(+ik\eta_0[a_{sf}^2-a_{si}^2]/2 ) } , \end{aligned}\ ] ] and that @xmath147/2 ) \over \sinh(\pi k \eta_0 a_{si}^2)\sinh(\pi k \eta_0 a_{sf}^2 ) } , \\ \sinh^2(\pi
k \eta_0 [ a_{sf}^2-a_{si}^2]/2 ) \over \sinh(\pi k \eta_0 a_{si}^2)\sinh(\pi
k \eta_0 a_{sf}^2 ) } .\end{aligned}\ ] ] these expressions are related to the bogoliubov coefficients by [ _ cf _ birrell - davies equation ( 3.93 ) ] @xmath148 @xmath149 the spectrum of particles in the final state is then @xmath150 which gives @xmath151 we now use the standard scattering theory result that a momentum - space delta function evaluated at zero is proportional to the volume of the `` universe '' , in this case the volume of the bec , @xmath152 to see that @xmath153 equivalently @xmath154 and the total number of emitted phonons is @xmath155 as a practical matter the integral will always be cut off at high momentum most typically by the inverse timescale @xmath156 over which the propagation speed changes , but if nothing else the integral can not be trusted for momenta higher than that associated with the healing length @xmath157 for the reasons previously discussed ( see sec .
[ sec : var - a ] ) .
in order to gain a better understanding of the particle creation just described , it may be useful to study separately the two opposite regimes characterizing this phenomenon .
in fact , for a given timescale of change , @xmath158 , driving the particle creation , one has a simplified description of the particle production when considering the case of modes with frequencies much smaller than @xmath159 ( sudden approximation ) or much larger than that ( adiabatic approximation ) .
after a brief discussion of these regimes we shall deal with the full intermediate case .
a particularly useful approximation is to take the `` sudden limit '' .
mathematically this consists of taking a step function for the scale factor @xmath160.\end{aligned}\ ] ] physically this means that one is considering that the change in @xmath137 is driven more rapidly than the frequency band one is interested in . in this case
, this means that the change is so fast that the entire acoustic regime is excited ( the time rate is trans - healing ) , but sufficiently slowly that the gp equation still continues to hold ( the time rate is still sub - interaction ) .
however we shall still have to put in `` by hand '' a high momentum cut off , given by the healing length @xmath161 , because beyond this point we can not trust the dispersion relation to remain on the acoustic branch implicit in our calculation .
the relevant calculation can be preformed by simply considering the mathematical @xmath162 limit in parker s model .
indeed the @xmath163 and @xmath164 coefficients are now momentum independent and @xmath165 @xmath166 as should be expected , particle production in this sudden limit depends only on the change in the scale factor .
the particle production spectrum is now flat ( more precisely , phase space limited ) all the way up to the healing frequency .
a rigorous result is that for _ any _ monotonic change in @xmath137 from @xmath131 to @xmath133 the magnitudes of the @xmath163 and @xmath164 coefficients are less than or equal to those calculated for the sudden approximation @xcite consequently the sudden approximation provides an absolute upper bound on particle production .
the number of phonons produced is @xmath167 that is @xmath168 the good news for current purposes is that this scales as @xmath169 . now
the trap size @xmath112 is of order 10 microns , while the healing length is in the range from 1 to 0.1 micrometers ; thus @xmath170 to @xmath171 .
a prefactor this big is desirable in terms of producing an observable effect . as for @xmath31 ( scattering length ) , this can range from 100 nm to 1 nm , so the ratio @xmath172 is up to @xmath173 .
since @xmath174 ( scale factor ) is @xmath175 we have @xmath176 up to @xmath177 .
therefore @xmath178 which is of order one
so there is no enormous suppression coming from the bogoliubov factor .
all in all , we estimate that @xmath179 to @xmath180 phonons can be produced in the sudden approximation .
in contrast , when the parameter @xmath181 is large compared with unity ( that is , for large enough momenta ) , and provided @xmath182 , we have @xmath183 this is a correctly normalized planckian [ black body ] distribution .
we can associate a temperature @xmath184 to the final phonon content produced by the expansion . before doing that ,
let us write down some useful relations between the frequencies associated with the pseudotime @xmath185 and with the laboratory time @xmath57 .
asymptotically ( either in the infinite past or in the infinite future ) the relation between the times @xmath57 and @xmath185 is given by @xmath186 so that @xmath187 here , the upper indices indicate with respect to which time a frequency is defined , and the sub - scripts @xmath188 and @xmath189 identify whether a particular magnitude is evaluated in the initial or in the final configuration .
now it is easy to see that the laboratory temperature that we would associate to the final configuration in the adiabatic approximation is @xmath190 in order to estimate this temperature we need to convert the time scale over which the scattering length changes , @xmath191 from the pseudo time to the laboratory time . to do this
we define @xmath192 for the particular profile , equation ( [ eq : prof ] ) , that associated with this model equation ( [ t - eta - rel ] ) evaluates to @xmath193 if we use this relation between @xmath194 and @xmath158 , the temperature would be @xmath195 in this way we check the intuitive idea that in the expansion process one would create phonons with frequencies inversely related with the temporal scale of change of the configuration in laboratory .
the number of phonons in the final state is @xmath196 contrary to what happens in the sudden approximation , the adiabatic approximation is taking into account the rapidity with which the configuration changes .
thus , the total number of phonons calculated by trusting the adiabatic approximation throughout the whole range of frequencies is suppressed with respect to the sudden approximation calculation by a factor of @xmath197 times @xmath198 ^ 3 $ ] this factor consisting of a dimensionless number coming from the detailed expression for the integral , times the cubed ratio of the healing time in the condensate to the time over which the scattering length is forced to change .
with @xmath199 we still get somewhere between @xmath200 and @xmath201 phonons .
once @xmath202 this is reduced to somewhere between @xmath203 and @xmath200 phonons ; but in this case we will also run into problems from finite volume effects @xmath194 is then comparable to the sound crossing time for the condensate and the momentum space delta functions appearing above are smeared out due to the finite volume of the condensate .
( this point is carefully addressed in a rather different physical context in @xcite , though many of the mathematical manipulations appearing therein are very similar to the present situation . )
note that as @xmath204 the adiabatic approximation calculation still results in one order of magnitude less than the equivalent calculation with the sudden approximation .
moreover , in this case , the whole range of observable phonons ( with frequencies between the healing frequency and the trap frequency ) is beyond the strict range of applicability of the adiabatic approximation ( remember that @xmath205 ) .
therefore , to be more precise one will have to make an intermediate analysis , in between the sudden and the adiabatic regimes .
let us estimate the value of the temperatures associated with the adiabatic approximation for temporal scales of change within the observable window . from equation ( [ temperature ] )
we can see that this temperature will be @xmath206 , and so , for @xmath194 between @xmath207 s ( associated with the condensate size ) and @xmath208 s ( the shortest timescale compatible with the acoustic approximation associated with the initial configuration ) will range from @xmath38 nk to @xmath209 nk . from this estimate
we can already see that , by modifying the scattering length on time scales close to the healing time , one could produce a bath of ( almost ) thermal phonons so energetic that even the mean field approximation might break down ( causing the complete disruption of the condensate ) .
when the peak frequency tends toward the healing frequency we see that the low frequency part of the observable spectrum will develop important departures from thermality .
we shall now turn to the general intermediate case and describe the full spectrum of phonons created in our toy model . then
, we will discuss the observability of the cosmological particle creation effect in terms of the ratio @xmath91 defined in equation ( [ c - ratio ] ) .
let us now consider an intermediate regime : look at large momenta and ignore for the time being any cutoff arising from the interaction timescale or the healing length , then @xmath210 \qquad \hbox{as } \qquad k \to \infty,\ ] ] so the particle spectrum is always exponentially suppressed at sufficiently high momenta . if we pick @xmath158 to be longer than a healing time ( so that we can not use the sudden approximation ) but still sufficiently short that we can not use the adiabatic approximation , then we will need to retain the full form of @xmath211 .
the resulting spectrum of phonons is @xmath212 \over \sinh[\pi \eta_0 k a_{sf}^2 ] \ ; \sinh[\pi \eta_0 k a_{si}^2 ] } , \label{eq : num - spec}\ ] ] and the total number of phonons produced is @xmath213 \over \sinh[\pi \eta_0 k a_{sf}^2 ] \ ; \sinh[\pi \eta_0 k a_{si}^2 ] } \;\d k,\ ] ] we can now consider the actual spectrum described by equation ( [ eq : num - spec ] ) by choosing plausible values for an experiment . however in order to get the spectrum that might be observed one has to convert the relevant quantities in equation ( [ eq : num - spec ] ) to the laboratory counterparts . using the expressions in equation ( [ omega - t - eta ] ) , and the relation between @xmath194 and @xmath158 given in equation ( [ t - eta - rel-2 ] ) , and alternatively rewritten here as @xmath214
the number spectrum can be written as ( we are reintroducing here the explicit _ in _ and _ out _ indices ) @xmath215 } { \sinh\left[2\pi \frac{\textstyle a^{4}_{sf } } { \textstyle a^{4}_{si}+a^{4}_{sf}}\omega^t_{out}t_0\right]\,\sinh\left[2\pi \frac{\textstyle a^{4}_{si } } { \textstyle a^{4}_{si}+a^{4}_{sf}}\omega^t_{in}t_0\right ] } \delta\left(\vec{k}_{in}+\vec{k}_{out}\right)\d^3\vec{k}_{in } , \label{eq : num - spec2}\ ] ] which implies @xmath216 } { \sinh\left[2\pi \frac{\textstyle a^{4}_{sf } } { \textstyle a^{4}_{si}+a^{4}_{sf } } \omega^t_{out } t_0\right]\ , \sinh\left[2\pi \frac{\textstyle a^{2}_{si}a^{2}_{sf } } { \textstyle a^{4}_{si}+a^{4}_{sf}}\omega^t_{out}t_0\right ] } % \left(\frac { ma_{sf}^{2 } } { n_{c } } \right)^{3 } \frac{(\omega^{t } _ { out})^2}{c_{sf}^3}. \label{eq : num - spec3}\ ] ] regarding the relative range of the scale factor @xmath130 , we have already seen that it can be deduced from the experimentally plausible range for the scattering length and we shall take @xmath217 . for the final value of the speed of sound @xmath218 we shall take @xmath219 mm / s
. in fact also the range of the speed of sound can be determined from the scattering length .
this can be reasonably be varied in the proximity of a feshbach resonance from 1 nm to 100 nm . for an experiment with heavy alkali atoms ( e.g. rubidium ) the speed of sound will then typically range from few mm / s to 10 mm / s .
the total number of phonons emitted is given by @xmath220 \over \sinh[x ( a_{sf}^2/a_{si}^2 ) ] \ ; \sinh[x ] } \;\d x = { 1 \over 16 \pi^5 } { v \over ( c_{sf}t_0)^3 } \left({a_{sf}^4+a_{si}^4 \over a_{sf}^2 a_{si}^2}\right)^3 f(a_{sf}^2/a_{si}^2),\ ] ] so qualitatively we have the same behaviour as in the adiabatic approximation , modulated by a dimensionless function of the ratio @xmath221 .
( this expression will remain valid as long as @xmath158 is longer than the healing time ; at which stage one should switch over to the sudden approximation . ) using hyperbolic trig identities the previous integral can be re - written as @xmath222 with @xmath223 for @xmath224 the function @xmath225 quickly and smoothly approaches its asymptotic value . in closing our analysis of the intermediate regime
we want now to compare the exact spectrum equation ( [ eq : num - spec3 ] ) with the spectrum obtained within the adiabatic approximation .
equation ( [ eq : spect - ad ] ) can be easily rewritten in the laboratory variables as @xmath226 - 1 } \frac{(\omega^{t } _ { out})^2}{c_{sf}^3}. \label{eq : num - spec - ad}\ ] ] the combined plot of ( [ eq : num - spec3 ] ) and ( [ eq : num - spec - ad ] ) is shown in figure [ fig : spectrum2 ] . it is evident that apart from a minor discrepancy at lower frequencies the two plots basically coincide from the peak frequency ( approximately 2 khz ) on .
this is not so surprising given that the adiabatic approximation implies @xmath227 which in laboratory variables corresponds to @xmath228 one can easily check that this inequality starts to be satisfied for frequencies of the order of a few khz ( and holds for any larger frequency ) . in this subsection
we will calculate the relevant ratio @xmath91 defined in equation ( [ c - ratio ] ) .
we want @xmath229 to make the phonons easily detectable . by using equation ( [ representation - change ] ) we can deduce that @xmath230 \rangle \right).\end{aligned}\ ] ] in the acoustic regime and for the particular configurations we are looking at we know that @xmath231 .
considering now that @xmath232=i \hbar \ ; \delta^3(x - y)$ ] or @xmath233=i \hbar \ ; \g \ ; \delta^3(x - y)$ ] we can deduce this starting from the fundamental commutation relation @xmath234=\delta^3(x - y)$]) we can arrive at @xmath235 to calculate this average , we can expand the real field operator @xmath48 in terms of the creation and annihilation operators associated with the final configuration @xmath236 and consider the appropriate particle content for the quantum state .
if @xmath237 is the average number of particles with momentum @xmath238 in the quantum state , the previous magnitude can be expressed as @xmath239 to reach this expression we have to substract the vacuum contribution . if we consider now the sudden approximation to calculate this rate ( an upper limit to what one could get ) we obtain @xmath240 with @xmath241 the one in equation ( [ n - sudden ] ) .
now @xmath242 ev .
instead @xmath243 ev @xmath244 ev . the factor @xmath245 , so we have that the relevant number @xmath246 . but @xmath247 and @xmath248 .
this gives @xmath249 .
this number is based on the sudden approximation and will be smaller in a more realistic calculation . however , remembering the discussion on the adiabatic approximation we know that , for temporal scales of change @xmath194 of the order of the healing time , the actual coefficient @xmath91 can not be smaller than @xmath197 times the previous estimate , i.e. @xmath250 .
therefore , by implementing in laboratory an expansive process with @xmath194 in an intermediate regime , in between the healing times @xmath251 associated with the initial configuration ( @xmath252 seconds ) and the final configuration ( @xmath253 seconds ) , one should be able to observe the effect .
in this work we have discussed the possibilities that becs offer for simulating lorentzian geometries of the cosmological type in the laboratory .
there are two inequivalent paths one can follow in this task . the first one is based on provoking an expansive explosion in the condensate by changing with time the characteristic frequency of an isotropic and harmonic confining potential .
this option implies that the velocity profile of the condensate acquires arbitrarily high values at large distances from the center .
so there is always a sphere at which the velocity of the expanding bec would surpass its sound velocity : a sonic horizon would be formed . in practice , due to the fact that physically realizable becs are finite systems , one can only reproduce on them a portion of an expanding universe .
therefore one might argue that the sonic horizon would be formed outside the bec .
however , the velocity of sound in becs is so small ( a few mm / s ) that in practical situations the sonic horizon will be formed well inside the system . now
the existence of sonic horizons is certainly interesting in its own right , but is not inherent to the simulation of cosmological spacetimes .
moreover , the plausible dynamical instabilities associated with their formation could mask the observation of purely cosmological effects .
the alternative path to simulating a cosmological geometry that we have pursued in this article is to take advantage of the possibility of varying the scattering length or , what is the same , the interaction strength between the atoms in the condensate . in this case
, what we need is a confining potential with a sufficiently large almost flat minimum in which a portion of the condensate stays at rest . in this configuration
, there is no formation of sonic horizons and thus we think it is ( both conceptually and technically ) a much cleaner path to follow .
the description of the condensation phenomenon naturally involves the separation of the system into a `` classical '' wave function ( the condensate part ) and quantum fluctuations . in the acoustic approximation
we can think of these quantum fluctuations as phonons over a classical background geometry , in this case , the analogue of a cosmological spacetime .
therefore , we can use the tuning of the scattering length to simulate not only a classical cosmologically expanding universe , but the quantum phenomenon of cosmological particle creation .
we have analyzed this well known process by using a minor variant of parker s model for a finite amount of expansion @xcite .
then , by working with numerical estimates appropriate to currently accessible becs in dilute gases , we have presented an analysis of the feasibility of observing the effect in real experiments .
we have seen that there is a more than plausible window for the observability of the effect with current technology . in current becs
the scattering length can easily be varied from a 100 nm to 1 nm .
this produces an expansion in the geometrical scale factor of about three times .
the temporal scale of change of the scattering length can not be arbitrarily short .
it has to be slower than the time scale in which the interaction between two atoms proceeds .
we have calculated this time scale to be of the order of microseconds .
however , we have also seen that , by driving the previous finite amount of expansion in a temporal scale of change of about fractions of millisecond , one could start to detect the presence of cosmological particle creation . from here
one could shorten the time scale down to tens of microsecond progressively amplifying the expected effect . by the time one reaches tens of microseconds the effect would have been amplified by a factor of a hundred , ( with timescale still above the interaction time ) , opening even the possibility of totally disrupting the condensate .
the relevant temporal scales of change for the effect to be observable are of the order of the healing time in the condensate .
therefore we expect that apart from the phonon spectrum calculated by neglecting the modified dispersion relations at high energies , there will be also some production of quasiparticles . to observe the purely cosmological effect one would have to keep this quasiparticle production under a certain level ; thus , the temporal scale of change should not be driven significantly beyond the healing time . in our analysis
we have neglected finite volume effects .
however we shall now show that these effects are insignificant for the typical condensate we considered here .
the fractional change in the number of particles produced due to finite volume effects is expected to be of order @xmath254=(cutoff wavelength)/(@xmath255size of the condensate ) .
the ratio between the healing length and the bec thomas fermi radius can be expressed as a ratio between the harmonic trap length and the scattering length .
@xmath256 for a harmonic oscillator length of about @xmath257 microns , @xmath258 atoms and a scattering length of 1 nanometer one gets @xmath259 . for a scattering length ten
times larger ( easily achievable with a feshbach resonance and still compatible with @xmath260 ) and @xmath261 m one would get @xmath262 .
this implies that @xmath263 and hence finite volume effects are negligible . to conclude ,
our analysis suggest that it should be already possible to observe the process of cosmological particle creation in bec analogue systems , by changing the scattering length from an initial value of about 100 nm to a final value of about 1 nm on times scales shorter than milliseconds but larger than tens of microseconds .
research by cb is supported by the ec under the marie curie contract hpmf - ct-2001 - 01203 .
research by sl is supported by the us nsf under grant no .
phy98 - 00967 and by the university of maryland .
research by mv is supported by the marsden fund administered by the royal society of new zealand [ rsnz ] .
sl wishes to thank t. jacobson for his help and support , b .-
hu , e. calzetta , k.burnett , y. castin , c. clark and w. phillips for illuminating discussions , and e. bolda , d. mattingly , g. pupillo and a. roura for their interesting remarks .
a. strominger and c. vafa , `` microscopic origin of the bekenstein - hawking entropy , '' phys .
b * 379 * , 99 ( 1996 ) [ arxiv : hep - th/9601029 ] .
b. l. hu and e. verdaguer , `` stochastic gravity : a primer with applications , '' class .
grav . * 20 * , r1 ( 2003 ) [ arxiv : gr - qc/0211090 ] .
b. l. hu and e. verdaguer , `` recent advances in stochastic gravity : theory and issues , '' arxiv : gr - qc/0110092 .
r. parentani , `` what did we learn from studying acoustic black holes ?
, '' int .
j. mod .
phys . a * 17
* , 2721 ( 2002 ) [ arxiv : gr - qc/0204079 ] .
w. g. unruh , `` notes on black hole evaporation , '' phys . rev .
d * 14 * , 870 ( 1976 )
. + w. g. unruh , `` experimental black hole evaporation ? '' , phys .
lett , * 46 * , 1351 ( 1981 ) .
m. visser , `` acoustic black holes : horizons , ergospheres , and hawking radiation , '' class .
grav . * 15 * ( 1998 ) 1767 [ arxiv : gr - qc/9712010 ] . j. martin and r. h. brandenberger , `` the trans - planckian problem of inflationary cosmology , '' phys .
d * 63 * , 123501 ( 2001 ) [ arxiv : hep - th/0005209 ] .
l. j. garay , j. r. anglin , j. i. cirac and p. zoller , `` black holes in bose - einstein condensates , '' phys .
lett . * 85 * , 4643 ( 2000 ) [ arxiv : gr - qc/0002015 ] .
l. j. garay , j. r. anglin , j. i. cirac and p. zoller , `` sonic black holes in dilute bose - einstein condensates , '' phys .
a * 63 * , 023611 ( 2001 ) [ arxiv : gr - qc/0005131 ] . c. barcel , s. liberati and m. visser , `` analog gravity from bose - einstein condensates , '' class .
quant . grav .
* 18 * ( 2001 ) 1137 [ arxiv : gr - qc/0011026 ] .
m. novello , m. visser and g. volovik , `` artificial black holes '' , ( world scientific , singapore , 2002 ) .
c. barcel , s. liberati and m. visser , `` towards the observation of hawking radiation in bose - einstein condensates , '' int .
j. mod .
phys . a , in press ; arxiv : gr - qc/0110036 . c. barcel , s. liberati and m. visser , `` analogue models for frw cosmologies '' , int .
j. mod .
d , in press ; arxiv : gr - qc/0305061 .
e. a. calzetta and b. l. hu , `` bose - novae as squeezing of the vacuum by condensate dynamics , '' arxiv : cond - mat/0208569 .
+ e. a. calzetta and b. l. hu , `` bose - novae as squeezing of vacuum fluctuations by condensate dynamics , '' arxiv : cond - mat/0207289 .
e. donley _
, nature * 412 * ( 2001 ) 295 ; + n. r. claussen , ph . d. thesis , u. of colorado ( 2003 ) .
s. l. cornish , n. r. claussen , j. l. roberts , e. a. cornell , c. e. wieman , `` stable 85rb bose - einstein condensates with widely tunable interactions '' phys .
lett * 85 * , 1795 ( 2000 ) .
arxiv : cond - mat/0004290 + j. l. roberts , n. r. claussen , s. l. cornish , e. a. donley , e. a. cornell , c. e. wieman , `` controlled collapse of a bose - einstein condensate '' arxiv : cond - mat/0102116 + elizabeth a. donley , neil r. claussen , simon l. cornish , jacob l. roberts , eric a. cornell , carl e. wieman , `` dynamics of collapsing and exploding bose - einstein condensates '' arxiv : cond - mat/0105019 p. o. fedichev and u. r. fischer , `` cosmological particle production in oscillating ultracold
bose gases : the role of dimensionality , '' arxiv : cond - mat/0303063 ; + `` hawking radiation from sonic de sitter horizons in expanding bose - einstein - condensed gases , '' arxiv : cond - mat/0304342 ; + `` observing quantum radiation from acoustic horizons in linearly expanding cigar - shaped bose - einstein condensates , '' arxiv : cond - mat/0307200 y. kagan , e. l. surkov and g. v. shlyapnikov , `` evolution of a bose - condensed gas under variations of the confining potential , '' phys . rev . *
a54 * , r1753 ( 1996 ) ; y. castin and r. dum , `` bose - einstein condensates in time dependent traps , '' phys .
. lett . * 77 * , 5315 ( 1996 ) .
y. kagan , e. l. surkov and g. v. shlyapnikov , `` evolution and global collapse of trapped bose condensates under variations of the scattering length , '' phys .
lett . * 79 * , 2604 ( 1997 ) [ arxiv : physics/9705005 ] .
volovik , `` induced gravity in superfluid 3he , '' j.low.temp.phys .
* 113 * , 667 ( 1997 ) [ arxiv : cond - mat/9806010 ] .
yvan castin , `` bose - einstein condensates in atomic gases : simple theoretical results '' in coherent atomic matter waves , lecture notes of les houches summer school , p.1 - 136 , edited by r. kaiser , c. westbrook , and f. david , edp sciences and springer - verlag ( 2001 ) .
m. visser , c. barcelo and s. liberati , `` acoustics in bose - einstein condensates as an example of broken lorentz symmetry , '' arxiv : hep - th/0109033 .
t. khler and k. burnett , `` microscopic quantum dynamics approach to the dilute condensed bose gas '' , phys . rev .
* a65 * , 033601 ( 2002 ) . c. j. williams , `` determination of the scattering lengths of @xmath264 from @xmath265 photoassociation line shapes '' , phys .
rev . * a60 * , 4427 ( 1999 ) .
e. l. bolda , e. tiesinga , and p. s. julienne , `` effective - scattering - length model of ultracold atomic collisions and feshbach resonances in tight harmonic traps '' phys .
rev . * a66 * , 013403 ( 2002 ) .
s. liberati , m. visser , f. belgiorno and d. w. sciama , `` sonoluminescence as a qed vacuum effect .
i : the physical scenario , '' phys .
d * 61 * , 085023 ( 2000 ) [ arxiv : quant - ph/9904013 ] .
s. liberati , m. visser , f. belgiorno and d. w. sciama , `` sonoluminescence as a qed vacuum effect .
ii : finite volume effects , '' phys .
d * 61 * ( 2000 ) 085024 [ arxiv : quant - ph/9905034 ] .
m. visser , `` some general bounds for 1-d scattering , '' phys .
a * 59 * ( 1999 ) 427 [ arxiv : quant - ph/9901030 ] . | bose - einstein condensates ( bec ) have recently been the subject of considerable study as possible analogue models of general relativity . in particular it was shown that the propagation of phase perturbations in a bec can , under certain conditions , closely mimic the dynamics of scalar quantum fields in curved spacetimes . in two previous articles [ gr - qc/0110036 , gr - qc/0305061 ] we noted that a varying scattering length in the bec corresponds to a varying speed of light in the `` effective metric '' .
recent experiments have indeed achieved a controlled tuning of the scattering length in rubidium 85 . in this article
we shall discuss the prospects for the use of this particular experimental effect to test some of the predictions of semiclassical quantum gravity , for instance , particle production in an expanding universe .
we stress that these effects are generally much larger than the hawking radiation expected from causal horizons , and so there are much better chances for their detection in the near future .
l s v l s v |
two decades after the pioneering work of allen and co - workers @xcite showing that laguerre - gaussian ( lg ) beams carry well defined orbital angular momentum ( oam ) , the role of this oam in interactions of such beams with an atom or a molecule remains an open question .
oam of light usually interacts with the external center - of - mass ( c.m . )
motion of an atom @xcite .
in contrast , light polarization which is spin angular momentum of light can interact with the internal electronic motion of an atom .
many researchers have predicted that the field oam can be transferred to the internal motion of an atom @xcite or a molecule @xcite in electronic dipole or quadrupole transitions , while some works @xcite have shown that the field oam does not interact with molecular chirality .
applications have been proposed based on direct coupling of field oam to the internal motion @xcite or to the c.m .
motion only @xcite .
the transfer of field oam to c.m .
motion of optically trapped particles ( optical spanner effect ) @xcite and bose - einstein condensates ( bec ) @xcite is well known .
but the experiments so far seem to contradict direct coupling of field oam with internal motion @xcite .
it is therefore important to understand how light oam takes part in light - atom or light - molecule interaction .
the oam of light is associated with the spatial inhomogeneity of field over the beam cross - section .
the question we address here : can an electron in an atom or a molecule feel the spatial variation of the field during its orbital motion ?
the paper is organized as follows . in sec .
ii , we have developed the theory corresponding to the method of oam exchange in interaction of lg beam with an atom or a molecule , separately .
iii presents numerical calculations of transition rates in interaction of atomic bec with lg beam as an example of our theory . finally , in sec .
iv , we have made some concluding remarks .
we consider a lg beam without any off - axis node propagating along @xmath0 axis of laboratory frame interacting with a cold atom or a cold molecule whose c.m .
wavefunction has extension as large as comparable to the wavelength of the light but smaller than the waist of the beam .
the atom or molecule experiences a local field of the type @xcite @xmath1,\end{aligned}\ ] ] where @xmath2 is the projection of @xmath3 on @xmath4 plane . @xmath5 and @xmath6 are the winding number and the waist of the beam , respectively ; and @xmath7 is the azimuth .
[ [ the - atom - radiation - interaction ] ] * the atom - radiation interaction : * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we consider the simplest atomic system composed of a nucleus of positive charge ( mass ) @xmath8 ( @xmath9 ) and an electron of negative charge @xmath10 ( @xmath11 ) . for simplicity
, the spin of the particles is ignored .
the center of mass coordinate of the atomic system is @xmath12 with @xmath13 being the total mass , @xmath14 and @xmath15 the coordinates of the electron and nucleus , respectively .
the hamiltonian of the atom - field system is @xmath16 where @xmath17 is unperturbed atomic hamiltonian and @xmath18 is the interaction hamiltonian derived in power - zineau - wooley ( pzw ) scheme @xcite .
@xmath19 is the electric polarization given by @xmath20 where relative coordinate ( internal ) @xmath21 .
the diameter of the region of lg beam of eq .
( 1 ) typically ranges between @xmath22 and @xmath23 m @xcite while the dimension of an electron orbital in an atom is of the order of a few angstrom .
this means @xmath24 and we therefore use taylor s expansion @xmath25 where @xmath26r\sin\theta + ( ik)\text{r}_{\text{c.m.}\perp}^{\vert l \vert } e^{il\phi_{\text{c.m.}}}e^{ikz_{\text{c.m.}}}r\cos\theta.\end{aligned}\ ] ] substituting eqs .
( 3 ) , ( 4 ) and ( 5 ) into eq . ( 2 ) , @xmath27 can be separated into dipole and quadrupole parts as given by @xmath28 @xmath29\sin\theta + ( ik)\text{r}_{\perp}^{\vert l \vert } e^{il\phi_{\text{c.m.}}}e^{ikz_{\text{c.m.}}}\cos\theta\right\rbrace + \text{h.c.}.\end{aligned}\ ] ] the dot products of the type @xmath30 are replaced by @xmath31 with @xmath32 and @xmath33 . in paraxial approximation , the @xmath34 component is negligible . equation ( 6 ) shows that within electric dipole approximation the polarization of the field interacts with the electronic motion and the field oam interacts only with the external c.m .
motion as also demonstrated by several authors @xcite . the first term in eq .
( 7 ) implies that in electric quadrupole transition , the field oam is coupled to the c.m .
motion only and the extra unit of angular momentum in electronic motion results from the quantized c.m .
motion of the atom as in ref .
the invariance of the interaction hamiltonian around the beam axis imposes the conservation of total angular momentum of the field plus atom system while the gradient of the field along radial direction couples the quantized c.m . and electronic motion .
this coupling of c.m . and
electronic motion is the main novel feature of this interaction .
our calculation shows that either of the terms @xmath35 or @xmath36 is nonzero depending on the sign of @xmath5 .
the quadrupole transition matrix element is given by @xmath37 , where @xmath38 denotes an unperturbed atomic state , i.e. , eigenstate of @xmath17 .
we assume @xmath39 , where the c.m .
wavefunction @xmath40 depends on the external potential that traps the atom and the internal electronic wavefunction @xmath41 can be considered to be a highly correlated coupled - cluster orbital @xcite .
@xmath42 where @xmath43 and @xmath44 stand for the average spatial width of @xmath40 and @xmath41 , respectively and @xmath45 . from eq .
( 8) we deduce the selection rule for magnetic quantum number @xmath46 .
the extra unit of angular momentum transferred to the electron comes from the quantized c.m .
motion of the atom .
the orientation of this angular momentum is the same as that of the field oam .
we have assumed that @xmath47 .
hence , the transition probabilities become insignificant for higher order multipole transitions unless the intensity of beam is very high . [
[ the - molecule - radiation - interaction ] ] * the molecule - radiation interaction : * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we consider a diatomic molecular ion , e.g. , h@xmath48 or hd@xmath49 for simplicity , comprising of three particles : two nuclei of mass ( charge ) @xmath50 ( @xmath8 ) and @xmath51 ( @xmath8 ) , and one electron @xmath11 ( @xmath10 ) as schematically shown in fig.1 .
a and b are two nuclei and @xmath52 is the center of mass of the molecule .
@xmath53 are the positions of the nuclei a and b with respect to the c.m .
, respectively .
@xmath14 is the position of the electron in c.m .
coordinate system .
a molecular axis is defined as position of a with respect to b , i.e. , @xmath54 .
the nuclei oscillations @xmath55 are written explicitly as @xmath56 , where @xmath57 is the equilibrium position of a with respect to b , @xmath58 is a constant vector with the same direction as that of @xmath59 and of magnitude given by the properties of the molecule . @xmath60 and @xmath61 are positions of the electron relative to the nuclei a and b , respectively . hence , @xmath62 , @xmath63 and @xmath64 , where @xmath65 and @xmath66 is unit vector along @xmath59 .
the unperturbed quantum state of the diatomic molecule is described by @xmath67 , where @xmath68 stands for the electronic state , @xmath69 stands for the vibrational wavefunction , @xmath70 stands for the rotational wavefunction , and @xmath40 stands for the c.m .
wavefunction .
the interaction hamiltonian is derived in pzw scheme as in the previous section .
again , we assume that molecular dimension is so small compared to the beam waist ( @xmath71 ) and use taylor s expansion about c.m .
position , @xmath72 where @xmath73 and @xmath74 is regular solid spherical harmonic function given as @xmath75 @xcite .
the normalization constant is @xmath76 . using the expression of @xmath14 and translation property of solid spherical harmonics @xcite , eq .
( 10 ) can be expressed as @xmath77\nonumber\\ & + ( ik)(1/2)\text{r}_{\text{c.m.}\perp}^{\vert l \vert } e^{il\phi_{\text{c.m.}}}e^ { ikz_{\text{c.m.}}}\left[\mathcal{r}_1^{0}(r { \bf{\hat{\textbf{r}}}})+\mathcal{r}_1^{0}(\textbf{r}_a+\textbf{r}_b)\right].\end{aligned}\ ] ] the second term of eq . ( 9 ) shows that one unit of angular momentum is transferred from external c.m .
motion to the internal rotational and electronic motion of the molecule .
if we retain only first two terms in eq .
( 9 ) and carry out the integration over @xmath78 , we get dipole interaction hamiltonian as given by @xmath79-\frac{l}{2}\text{r}_{\text{c.m.}\perp}^{\vert l \vert -1}e^{\text{sgn}(l)i(\vert l\vert -1)\phi_{\text{c.m.}}}e^{ikz_{\text{c.m.}}}\left[ry_1^{\sigma}({\bf{\hat{\textbf{r}}}})\mathcal{r}_1^{\text{sgn}(l)1}(\textbf{r}_a+\textbf{r}_b)\right.\nonumber\\ & \left.\left.+(r_ay_1^{\sigma}({\bf\hat{\textbf{r}}}_a)+r_by_1^{\sigma}({\bf\hat{\textbf{r}}}_b))\mathcal{r}_1^{\text{sgn}(l)1}(r { \bf{\hat{\textbf{r}}}})\right]\right\rbrace + \text{h.c.}.\end{aligned}\ ] ] the transition matrix element of the dipole interaction hamiltonian ( 12 ) is @xmath80 \mathcal{m}_{\text{c.m.}}^{l } \nonumber\\ & -\frac{l}{2}\sqrt{\frac{2\pi } { 3}}\left(\dfrac{m_2-m_1}{m_2+m_1}\right)\left(\dfrac{w_{\text{c.m.}}}{w_0}\right)^{\vert l\vert -1}\left(\dfrac{\bar{r}}{w_0}\right)\left[\langle \psi_{ef}\vert ( r_ay_1^{\text{sgn}(l)1}({\bf\hat{\textbf{r}}}_a)+r_by_1^{\text{sgn}(l)1}({\bf\hat{\textbf{r}}}_b))\vert \psi_{ei}\rangle \mathcal{m}_{r1}^{\sigma } \mathcal{m}_{v}^1 \right.\nonumber\\ & \left.\left.+\langle \psi_{ef}\vert ( r_ay_1^{\sigma}({\bf\hat{\textbf{r}}}_a)+r_by_1^{\sigma}({\bf\hat{\textbf{r}}}_b))\vert \psi_{ei } \rangle \mathcal{m}_{r1}^{\text{sgn}(l)1 } \mathcal{m}_{v}^1 \right]\mathcal{m}_{\text{c.m.}}^{\text{sgn}(l)(\vert l\vert -1)}\right\rbrace,\end{aligned}\ ] ] where @xmath45 , @xmath81 and @xmath82 . it is clear from eq .
( 12 ) and ( 13 ) that the c.m .
motion can couple with the internal motion in interaction of lg beam with diatomic molecule even within electric dipole approximation . in case of homonuclear ( non - polar )
molecules , @xmath83 .
( 12 ) clearly shows that rotational and vibrational motion of non - polar molecules can not be influenced by oam of lg beam .
however , when non - polar diatomic molecules are polarized and adiabatically or non - adiabatically aligned with a non - resonant intense laser field @xcite , then it is possible for the aligned molecules to interact with a lg beam , leading to transfer of light oam to the rotational motion of the molecules as we elaborate below .
particularly interesting prospect for light oam transfer to molecular electronic motion opens if one makes use of the rotational confinement due to the alignment of internuclear axes of diatomic molecules with a linearly polarized off - resonant intense laser field @xcite .
the alignment and orientation of diatomic molecules and consequent phenomena of anisotropic ionization @xcite and high harmonic generation ( hhg ) @xcite , have now become an important topic of research both from experimental and theoretical points of view , providing new insight into the underlying effects of rotational confinement on molecular electronic orbitals . since
, the rotation of internuclear axis of a diatomic molecule can couple to the molecular electronic orbital , the interaction of aligned molecules with light carrying oam will be interesting for oam transfer in dipole interactions .
the alignment of nonpolar ground state molecules is basically a result of two - photon off - resonant dipole interaction of molecules .
the major advantage of using pre - aligned molecules would be to get a preferential direction of orientation for angular momentum transfer .
if a far - off - resonant and relatively intense lg beam is allowed to interact with pre - aligned molecules , it is possible to orient the axis of the molecules by the transfer of light orbital angular momentum .
since the molecular axis orientation is coupled to the electronic angular momentum , light oam will eventually alter the electronic angular momentum .
thus , our theoretically proposed mechanism of light oam transfer can be experimentally verified by applying an lg beam to diatomic molecules which are already aligned and subsequently probing rotational dynamics with a probe laser via detecting ionization and hhg signals @xcite .
the light oam transfer can be inferred by comparing results with and without lg beam .
thus , with the currently available technology of molecular alignment and orientation with pump - probe type setup , it is possible to manipulate molecular rotational and electronic orbital motion with light carrying orbital angular momentum . in the next section ,
we discuss numerical results showing the effect of the coupling of c.m .
and electronic motions on the quadrupole transition rate in interaction of a lg beam with @xmath84na bec .
in the previous section , we have shown that the quadrupole transition is the lowest order transition in interaction of cold atoms with an lg beam for transferring field oam to electronic motion via quantized c.m . motion . here
we present numerical results showing the effect of the coupling between c.m .
and internal motion on quadrupole transition rates in interaction of an atomic bec with an lg beam . the first term of eq .
( 8) shows this transfer of optical oam to electronic motion via quantized c.m . motion .
this term is important to reveal the predicted effect in quadrupole transitions .
here we calculate the quadrupole transition rates considering quantum mechanical and coupled motion of both electronic and c.m .
degrees of freedom of atoms .
we consider the c.m .
atomic state as the ground state and different vortex states of a trapped bec @xcite .
the system is dilute enough so that the gross - pitaevskii ( gp ) theory @xcite for trapped bosons is applicable to the c.m .
vortex states in bec will be created due to transfer of oam from the lg beam to the c.m .
motion of the atoms .
in addition to the oam , linear momentum ( lm ) of light will also be transferred to the c.m .
motion of atoms .
hence , we write the initial and final stationary state of c.m . motion of atoms as @xcite @xmath85 and @xmath86 where @xmath87 is the quantum of circulation of atoms about @xmath0 axis .
@xmath88 represents vortex states of the bec .
the expression of @xmath89 gives the selection rule for c.m .
motion as @xmath90 . according to time - independent gp theory , the nonlinear schrdinger equation obeyed by the c.m .
motion in an anisotropic simple harmonic potential trap is given by @xcite @xmath91\psi_{\text{c.m.}}=\mu \psi_{\text{c.m.}}\end{aligned}\ ] ] where @xmath92 and @xmath93 are the two angular frequencies associated with the external potential of the anisotropic trap , @xmath94 is the @xmath95-wave scattering length , and @xmath96 is the chemical potential . in eq .
( 8) , @xmath43 is of the order of the characteristic length of trap @xmath97 and @xmath44 is of the order of bohr radius @xmath98 .
we evaluate the c.m .
wavefunction at zero temperature using the steepest descent method for functional minimization as prescribed in ref .
the electronic portion of the transition matrix element is calculated using coupled - cluster theory @xcite .
we now proceed to numerically evaluate the quadrupole transition rates considering the first term of eq .
( 8) where the c.m . and
electronic motions are coupled .
let us consider a left circularly polarized lg beam ( @xmath99 in eqs .
( 6 - 8 ) ) with @xmath100 is interacting with a bec of @xmath101 number of @xmath84na atoms in an anisotropic harmonic trap .
the axis of the beam and the axis of the trap are same and along z axis of the laboratory frame . here
it is important to note that , one unit of the field oam changes the vorticity of the c.m .
wavefunction and the other unit of field oam is transferred to electronic motion via quantized c.m .
motion . for numerical illustration
, we choose the characteristics of the experimental trap as given in ref .
the asymmetry parameter of the trap is @xmath102 . the axial frequency @xmath103 hz . the corresponding characteristic length is @xmath104 m. @xmath95- wave scattering length @xmath105 nm @xcite .
the waist of the lg beam is @xmath106 m and the intensity @xmath107 w@xmath108 .
the amplitude of the lg beam in eq .
( 8) is related to the intensity by @xmath109 where @xmath110 is vacuum permittivity .
the initial electronic state is @xmath111 . the electronic portion @xmath112 of the second term in the r.h.s . of eq .
( 8) indicates that the magnetic quantum number of final electronic state @xmath113 will be changed by one unit due to the polarization of light .
so , the final electronic state will be @xmath114 or @xmath115 .
but the electronic portion @xmath116 of the first term on the r.h.s . of eq .
( 8) shows that , in addition to the polarization of the beam , one unit of field oam can be transferred to the electronic motion via quantized c.m . motion .
that increases the magnetic quantum number by two units in the final electronic state , i.e. , @xmath117 .
thus , the electronic transition @xmath118 can be realized .
next , we show how quadrupole transition rates are enhanced due to quantized c.m .
motion and its coupling with the internal motion .
2 . presents the radial part of the c.m .
wavefunction of the cold atom corresponding to the ground state and different vortex states in the simple harmonic potential trap .
the value of the electronic portion of the transition matrix element remains unchanged during variation of vorticity of the c.m .
wavefunction . in ground state of c.m .
motion , the probability of finding the atom on the beam axis is maximum .
but as the vorticity of the states increases the probability of finding the atom away from beam axis increases . hence , the value of the c.m .
matrix element @xmath119 increases .
this in turn influences the electronic motion due to coupling with c.m .
table i presents variation of the c.m .
transition matrix element and the quadrupole transition rates ( @xmath120 ) with vorticity of involved c.m . states .
it is clearly shown that with increase in @xmath121 the quadrupole transition rate increases due to increase in spread of the c.m .
it is worth pointing out that @xmath122 is enhanced because the c.m .
motion is treated quantum mechanically and it is coupled with the electronic motion .
.variation of @xmath119 ( in unit of @xmath123 ) and quadrupole transition rates ( in @xmath101s@xmath124 ) with vorticity of initial state of c.m .
wavefunction . [ cols="^,^,^,^,^,^,^",options="header " , ] before ending this section , it is worthwhile to mention that , though for simplicity we have neglected electron s spin , the theoretical treatment we have presented can be extended to include electron s spin , in particular in situations where spin - orbit interaction of atoms or molecules is strong . in that case and in absence of hyperfine interaction ,
the good quantum number of an atom or molecule will be @xmath125 , where @xmath126 and @xmath127 stand for total orbital and spin angular momentum of all the valence electrons of an atom or a molecule .
in case of diatomic molecules , the projection @xmath128 of @xmath129 on the internuclear axis of the molecule will be a good quantum number .
as in atoms and molecules , spin - orbit interaction also arises in light .
circular polarization of light is associated with the spin of a photon . at a fundamental level , it is the vectorial sum of polarization and orbital angular momenta of photon which is a conserved quantity .
light oam will be nearly a good quantum number under paraxial approximation only . in recent times several experiments
@xcite have demonstrated the transfer of angular momentum in light - matter interactions through spin - orbit interactions or inter - conversion between spin and angular momenta of photons in different physical systems , for example , in nematic liquid crystal @xcite and nanoplasmonics @xcite .
the transfer of angular momentum in interaction of light with nematic liquid crystal is quite interesting .
the mechanism of light orbital angular momentum transfer presented in this paper is consistent with the method of oam transfer in liquid crystal as described in ref.@xcite .
in conclusion , we have studied the interaction of lg beam with an atom and a diatomic molecule .
we have found that if the atom or molecule is cold enough for its c.m .
motion to be quantized , then an angular momentum exchange can take place between c.m . and internal motion of the system .
atomic and molecular dimensions are far too small compared to the core size of the lg beam . however
, if the atom or molecule is cooled down to its recoil limit such that the spread of its c.m . wavefunction is comparable to the wavelength of the beam , then it can feel the spatial variation of the electric field along the radial direction .
the interaction with lg beam of eq .
( 1 ) ensures that the orientation of the angular momentum transferred from c.m . to internal motion
is same as that of the field oam .
our calculations clearly show that the extra angular momentum ( other than that coming from the polarization of the field ) to internal motion is coming from the quantized c.m . motion
. it may be possible to observe this effect with ultra - cold atoms or molecules interacting with lg beam as suggested by van enk @xcite twenty years ago .
we have numerically calculated the quadrupole transition rates in interaction of lg beam with atomic bec where both the c.m . and
electronic motions are quantized and coupled .
we have shown the dependence of the transition rates on the vorticity of the c.m .
wavefunction involved in the transition . in case of molecules at ultracold temperature or in quantum degenerate molecular gases , it would be possible to transfer the field oam to internal motion even in dipole interaction .
one potential application of the discussed effect can be thought of in quantum information processing using entangled angular momentum observables belonging to the same atom or molecule .
muthukrishnan and stroud @xcite have shown the entanglement between electronic and c.m .
degrees of freedom of a cold atom interacting with lg beam .
they have considered electric dipole interaction where the field oam couples to c.m . only and the field polarization couples to electronic motion . if the c.m .
motion is allowed to couple with the internal degrees of freedom through laser - generated synthetic gauge field @xcite then it is possible to realize transfer of light orbital angular momentum into the internal motion in otherwise similar experimental scenario of ref .
this will provide a new avenue for entanglement manipulation in angular degrees - of - freedom . in this context
, it is encouraging to note that the entanglement between electronic and vibrational degrees - of - freedom is already produced for an ion in a linear trap @xcite which can be a suitable system to explore the effect predicted in this paper .
our calculations further suggest that the quantized c.m .
motion of an atom can couple to electronic motion in electric quadrupole interaction .
the situation is even more interesting in case of molecule .
( 12 ) suggests that the c.m . , electronic , rotational and vibrational motion can be coupled in interaction with lg beam even in electric dipole approximation .
we have discussed how the induced polarization and alignment of diatomic molecules due to linearly polarized intense laser field can be useful for experimental verification of our theory . the alignment and orientation of the molecules of nematic liquid crystal have been found to play an important role in oam transfer @xcite , indicating that similar effects in gas phase molecules in intense laser fields will be important for light oam transfer to the molecules . finally , the recent advent of ultra long - range cold rydberg molecules whose electronic orbital can be as large as 100 nm @xcite and the demonstration of the alignment of such molecules @xcite open a new avenue for studying oam transfer in light - molecule interactions with huge enhancement of the size effects for the ability of the molecular electrons to experience the spatial variation of light intensity . | exchange of orbital angular momentum between laguerre - gaussian beam of light and center - of - mass motion of an atom or molecule is well known .
we show that orbital angular momentum of light can also be transferred to the internal electronic or rotational motion of an atom or a molecule provided the internal and center - of - mass motions are coupled .
however , this transfer does not happen directly to the internal motion , but via center - of - mass motion .
if atoms or molecules are cooled down to recoil limit then an exchange of angular momentum between the quantized center - of - mass motion and the internal motion is possible during interaction of cold atoms or molecules with laguerre - gaussian beam .
the orientation of the exchanged angular momentum is determined by the sign of the winding number of laguerre - gaussian beam .
we have presented selective results of numerical calculations for the quadrupole transition rates in interaction of laguerre - gaussian beam with an atomic bose - einstein condensate to illustrate the underlying mechanism of light orbital angular momentum transfer .
we discuss how the alignment of diatomic molecules will facilitate to explore the effects of light orbital angular momentum on electronic motion of molecules . |
the top quark is the heaviest fermion in the standard model .
its mass is at the electroweak symmetry - breaking scale . due to its large mass
its weak decay time is much shorter than the typical time for the strong interactions to affect its spin @xcite .
therefore the information on its polarization , is not disturbed by hadronization effects but transferred to the decay products . within the standard model ,
the dominant decay of the top quark is @xmath7 .
the angular distributions of the top quark decay products are determined by the momentum and spin state of the top quark itself .
it is believed that top quark physics will be crucial to understand the structure of standard model and existence of new physics beyond .
detailed discussions on top quark spin polarization properties will contribute future researches on this subject . in this work
we investigate top quark spin polarization along the direction of various spin bases for the single production in @xmath8 collision via the process @xmath9 .
the research and development on linear @xmath3 colliders have been progressing and the physics potential of these future machines is under study .
after linear colliders are constructed its operating modes of @xmath0 and @xmath10 are expected to be designed @xcite .
real gamma beam is obtained through compton backscattering of laser light off linear electron beam where most of the photons are produced at the high energy region .
the luminosities for @xmath0 and @xmath10 collisions turn out to be of the same order as the one for @xmath3 @xcite , so the cross sections for photoproduction processes with real photons are considerably larger than virtual photon case . in our calculations
we consider three different center of mass energies @xmath2=0.5 , 1 and 1.5 tev of the parental linear @xmath3 collider .
there are many detailed discussions in the literature for single top quark production and spin correlations in @xmath5 , @xmath11 and @xmath6 collisions @xcite . at @xmath5 and @xmath11 colliders
top quarks are produced @xmath12 and single t production modes . in double production case ,
spin up top quark and spin down anti - top quark are more likely produced and there are observable angular correlations among the decay products .
@xmath13 decay products in the final states give larger statistics than the single production case . on the other hand , final state in the single top case
is relatively simple and may compensate for the smaller statistics @xcite . in @xmath6 collisions , top quark decay products in the final states are dominated by the single top production due to absence of the @xmath12 production . since single top production contains an electroweak process with produced top quarks coupled to a w boson , decay products of the final top quark gives a high angular correlation @xcite .
this is also the case in the @xmath8 collision .
moreover linear @xmath3 collider or its @xmath0 mode provide a clean environment to study polarization phenomena of top quarks .
in @xmath3 linear colliders a hard photon beam can be produced by compton backscattering of laser light off linear electron or positron beam .
we consider the case in which the photon beam is obtained by compton backscattering off linear electron beam and the positron beam directly takes part in the subprocess .
the spectrum of the backscattered photons is given by @xcite .
@xmath14\end{aligned}\ ] ] where , @xmath15 with @xmath16 .
@xmath17 is the energy of initial laser photon and @xmath18 is the energy of initial electron beam before compton backscattering .
@xmath19 is the fraction which represents the ratio between the scattered photon and initial electron energy for the backscattered photons moving along the initial electron direction .
maximum value of @xmath19 reaches 0.83 when @xmath20 in which the backscattered photon energy is maximized without spoiling the luminosity .
the integrated cross section over the backscattered photon spectrum is given by : @xmath21 where @xmath22 and @xmath23 is the square of the center of mass energy of the subprocess @xmath9 .
@xmath23 is related to @xmath24 , the square of the center of mass energy of @xmath3 by @xmath25 .
the single production of top quark via the process @xmath1 is described by four tree level diagrams .
each of the diagrams contains wtb vertex and due to its v - a structure , produced top quarks are highly polarized . because of its large mass the helicity of a top quark is frame dependent and changes under a boost from one frame to an other .
the helicity and chirality states do not coincide with each other and there is no reason to believe that the helicity basis will give the best description of the spin of top quarks . so it is reasonable to study other spin bases better than helicity for top quark spin .
the spin four - vector of a top quark is defined by @xmath26 where @xmath27 in the top quark rest frame .
the laboratory frame is the @xmath3 center of mass system where the cross section is performed .
we consider four different top spin direction in the laboratory frame ; the incoming positron beam , photon beam directions and outgoing @xmath28 direction and also the helicity basis .
@xmath29 should be obtained by a lorentz boost from laboratory frame : @xmath30 here @xmath31 is the momentum of the particle moves along the top spin direction in the laboratory frame and @xmath32 is the momentum observed in the rest frame of the top quark . in table
i - iii polarized cross sections , dominant spin fractions and spin asymmetries are given for the various top quark spin bases .
it is shown from these tables that @xmath4-beam direction gives the highest degree of polarization ; @xmath33 at @xmath34 tev , @xmath35 at @xmath36 tev and @xmath37 at @xmath38 tev .
its energy dependence is significant . in relatively low energies
, helicity will not give the best description of the spin of a massive fermion like top quark .
therefore other spin bases may give a better description of top quark spin .
in our case favorite spin basis is the @xmath4-beam spin decomposition axis .
when energy increases top quark gradually becomes ultrarelativistic and spin fraction in the helicity basis grows .
this behaviour is shown from table i - iii .
spin fraction for the helicity basis increases with energy ; @xmath39 at @xmath34 tev , @xmath40 at @xmath36 tev and @xmath41 at @xmath38 tev . at the energy region
we have considered speed of the top quarks are not ultrarelativistic therefore helicity is not the favorite spin basis .
it is straightforward to obtain similar results for anti top quarks with the process @xmath42 . in this case
favorite spin basis is the @xmath43-beam spin decomposition axis .
same results at table i - iii for the spin fractions are obtained with the interchange of the bases ; @xmath4-beam @xmath44 @xmath43-beam , @xmath28-beam @xmath44 @xmath45-beam .
but the spin orientations should be reversed ; spin up @xmath44 spin down , l @xmath44 r. another useful quantity about spin - induced angular correlations is the spin asymmetry @xmath46 here subscript up arrow @xmath47 ( down arrow @xmath48 ) stands for spin up @xmath49 ( spin down @xmath50 ) and @xmath51 represents number of event for the corresponding spin .
the angular distribution of the top quark decay involves correlations between top decay products and top quark spin : @xmath52 here the dominant decay chain of the top quark in the standard model @xmath53 is considered .
@xmath54 is defined as the angle between top quark decay products and the top quark spin quantization axis in the rest frame of the top quark .
@xmath55 is the correlation coefficient and @xmath56 for @xmath57 or @xmath58 which leads to the strongest correlation .
one can see from table i that @xmath4-beam basis improves the asymmetry a factor of 4.35 when compared to helicity basis at @xmath34 tev . from table ii and iii
we see that this factor takes the value of 2.38 and 1.85 respectively . in order to get an idea about the influence of spin polarization on transverse momentum @xmath59 distributions of singly produced top quarks
we plot fig i - iii .
similar features in the tables are reflected in the figures also ; deviations of the @xmath59 distributions from the unpolarized ( total ) is the largest for helicity basis . in our calculations
phase space integrations have been performed by grace @xcite which uses a monte carlo routine .
we have shown that in the energy region @xmath60 tev @xmath4-beam direction provides a high degree of polarization .
considerable improvements have been obtained in the spin fractions and asymmetries with respect to helicity basis .
therefore a detailed analysis of top spin polarization and studying various spin bases in the @xmath8 collision is important and may contribute future researches .
linear @xmath3 collider and its @xmath0 mode provide a clean environment and the experimental clearness is an advantage of @xmath0 collisions with respect to @xmath5 , @xmath11 and @xmath6 collisions . at hadron
colliders the interacting particles are not the beam particles themselves .
the reactions are initiated by one of many partons present in the incident hadron . so the energy and quantum state of the initial state are not fixed . on the other hand lepton colliders
tend have much cleaner beams and lower backgrounds .
99 i. bigi , y. dokshitzer , v. khoze , j. kuhn and p. zerwas , phys . lett . * b 181*,157 ( 1986 ) .
c. akerlof , ann arbor report no .
um he 81 - 59 ( 1981 ) .
barklow , in proceedings of the 1990 summer study on research directions for the decade ( snowmass , colorado , 1990 ) , and slac report no .
slac - pub-5364 ( 1990 ) .
i.f . ginzburg _
et al . _ ,
methods * 205 * , 47 ( 1983 ) ; _ ibid . _ * 219 * , 5 ( 1984 ) .
g. mahlon and s. parke , phys .
* b 476 * , 323 ( 2000 ) ; + t.m.p .
tait , phys rev . *
d 61 * , 034001 ( 2000 ) ; + a.s .
belyaev , e.e .
boos and l.v .
dudko , _ ibid . _
* d 59 * , 075001 ( 1999 ) ; + t. stelzer , z. sullivan and s. willenbrock , _ ibid .
_ * d 58 * , 094021 ( 1998 ) ; + g. mahlon and s. parke , phys . lett .
* b 411 * , 173 ( 1997 ) ; + a.p .
heinson , a.s .
belyaev and e.e .
boos , phys rev * d 56 * , 3114 ( 1997 ) ; + mahlon and s. parke _ ibid .
55 _ , 7249 ( 1997 ) ; + s. parke and y. shadmi , phys . lett .
b 387 * , 199 ( 1996 ) . g. mahlon , talk presented at the thinkshop2 on top - quark physics for run ii ( fnal , 2000 ) and hep - ph/0011349 .
ata and b.d .
ahin , phys . rev .
* d 70 * , 037503 ( 2004 ) t. kaneko in `` new computing techniques in physics research '' , ed .
d. perret - gallix , w. wojcik , edition du cnrs , 1990 ; + minami - tateya group , `` grace manual '' , kek report 92 - 19 , 1993 ; + f. yuasa _ et al .
138 ( 2000 ) 18 .
distributions of the singly produced top quarks via process @xmath61 at center of mass energy @xmath34 tev of the parental linear @xmath3 collider .
dominant spin bases @xmath62-beam down , @xmath4-beam up , helicity left , @xmath28-beam down and unpolarized ( total ) case are drawn .
[ fig1 ] ] .polarized cross sections , dominant spin fractions and spin asymmetries for the various top quark spin bases in the production of single top process @xmath61 at center of mass energy @xmath34 tev of the parental linear @xmath3 collider .
[ tab1 ] [ cols="^,^,^,^",options="header " , ] | we investigate the degree of spin polarization of single top quarks in the @xmath0 collision via the process @xmath1 with center of mass energies @xmath2=0.5 , 1 and 1.5 tev of the parental linear @xmath3 collider .
dominant spin fractions and spin asymmetries for the various top quark spin bases are investigated .
we show that @xmath4-beam direction is the favorite top quark spin decomposition axis .
it is found to be comparable with the ones in @xmath5 and @xmath6 collisions .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ preprint of this article submitted for consideration in modern physics letters a ( 2007 ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
let @xmath4 be a number field and @xmath7 an elliptic curve .
an @xmath3-cycle of @xmath7 is a cyclic subgroup of @xmath8 of order @xmath3 which is invariant under the action of gal(@xmath9 ) .
an @xmath3-cycle @xmath10 of @xmath7 gives rise to a curve @xmath11 and an isogeny @xmath12 defined over @xmath4 with kernel cyclic of order @xmath3 , and every such isogeny arises this way @xcite*rmk 4.13.2 .
if @xmath7 has an isogeny of this form we say @xmath7 has an @xmath3-isogeny .
if the points of an @xmath3-cycle are rational over an extension @xmath13 , we will say the corresponding @xmath3-isogeny is _ pointwise rational _ over @xmath14 .
let @xmath2 denote the cyclic group of order @xmath3 . in this paper , we classify @xmath3-isogenies defined over @xmath15 or @xmath1 that are pointwise rational over quadratic extensions of @xmath4 for @xmath16 @xmath17 ( with one exception)([isogenies ] ) . in the case
@xmath18 , @xmath5=c_1 $ ] we determine ( i ) a classification of the torsion structures which occur as the quadratic twists of a given torsion structure , ( ii ) a classification of the torsion structures which occur as the growths of a given torsion structure and ( iii ) tight bounds on the number of quadratic extensions in which a given torsion structure can grow ( [ oddtwist ] ) . in the case
@xmath19 we did not complete the classification because we could not disprove the existence of a 21-isogeny over @xmath4 .
this was accomplished in @xcite*prop .
2 . for a history of related classification problems
please see @xcite .
there is an affine curve @xmath20 whose @xmath4-rational points classify isomorphism classes of pairs @xmath21 where @xmath7 is an elliptic curve and @xmath10 is an @xmath3-cycle .
two pairs @xmath21,@xmath22 are equivalent if and only if there is an isomorphism @xmath23 such that @xmath24 . by adding a finite number of points ( called @xmath25 ) to @xmath20 we
obtain the projective curve @xmath26 .
the curve @xmath26 has a model over @xmath27 and hence we have tools to study @xmath28 .
let @xmath7 be an elliptic curve and let @xmath13 be a quadratic extension .
we summarize our strategy as follows : when @xmath5 = c_1 $ ] , we have @xmath29 = c_1 $ ] ( [ gt ] ) . hence in order to complete tasks ( i ) , ( ii ) and ( iii ) , it suffices to complete ( i ) by [ oddpartsadd ] .
but by [ twistback ] , if @xmath30 and @xmath31 then one can often show @xmath32 has an @xmath3-isogeny pointwise rational over @xmath33 for some large value of @xmath3 , and these are rare over @xmath4 ( [ isogenies ] ) .
the classification of @xmath3-isogenies leads to the diophantine problem of determining @xmath28 . in section 2
we describe some results necessary to understand the rest of the paper . in section 3
we study the @xmath4-rational points on @xmath26 for @xmath34 , @xmath35 and @xmath36 .
in section 4 we use the classification of @xmath3-isogenies to study growth of torsion .
computation played an important role in our work .
we used magma to compute the rank and torsion of elliptic curves over number fields .
we also used magma to find automorphism groups of curves and compute quotient curves under the action of certain groups .
the classification of @xmath3-isogenies relied upon the small modular curves package in magma .
we require the following classification theorem .
( najman @xcite ) [ najman ] let k be a cyclotomic quadratic field and e an elliptic curve over k. * if @xmath37 then @xmath38 is either one of the groups from mazur s theorem @xcite*thm 2 or @xmath39 . * if @xmath0 then @xmath38 is either one of the groups from mazur s theorem , @xmath40 or @xmath41 .
4 [ oddpartsadd ] if n is an odd positive integer we have @xmath42 \approx e(k)[n ] \oplus e^d(k)[n]\ ] ] [ twistback ] let k be a number field and e / k an elliptic curve .
let d @xmath43 k be a nonsquare and let l = k@xmath44 . if h is a subgroup of @xmath45 of odd order , then there is a gal(@xmath9)-invariant subgroup j of e(l)@xmath46 such that j @xmath47 .
we may assume e is in weierstrass form .
we have an isomorphism : @xmath48
@xmath49 so @xmath50 and @xmath51 is gal(@xmath9)-invariant since the points of @xmath52 are rational over @xmath4 and @xmath52 is a subgroup ( so closed under inverses ) .
since @xmath52 has odd order , it has no points of order @xmath53 so @xmath54 .
hence @xmath55 is a subgroup of @xmath56 . as @xmath57 is the sum of gal@xmath58-invariant subgroups , it is invariant as well .
[ full ] let k@xmath59 ( @xmath60 ) , e / k an elliptic curve and l a quadratic extension of k. then the only odd prime power n such that @xmath61 is n=3 .
let @xmath62 denote euler s totient function .
if @xmath63 where @xmath64 is a prime and @xmath61 then by @xcite*cor 8.1.1 , @xmath14 contains an @xmath3th root of unity @xmath65 .
hence @xmath66 = 4 $ ] so @xmath67 or @xmath68 .
note there is either a 3rd or 4th root of unity in @xmath4 , so if @xmath69 is in @xmath14 , then there is a 15th or 20th root of unity in @xmath14 . but
@xmath70 and @xmath71 , a contradiction . on the other hand
, there is an elliptic curve e/@xmath27 ( namely @xmath72 $ ] ) which has full 3-torsion over k@xmath73 and hence provides examples in each case with l@xmath74 .
the following theorem lists various restrictions on growth in quadratic extensions .
[ gt ] let k be a number field , e / k an elliptic curve , l a quadratic extension of k and p an odd prime . 1 . if e(k)[2 ] @xmath75 then e(l)[2 ] @xmath75 .
[ gt 1 ] 2 .
if d @xmath43 k , d @xmath76 , then @xmath77 \approx e(k)[2]$ ] .
if e(k)[p ] @xmath75 and e(l)[p ] @xmath78 then k contains a primitive pth root of unity .
4 . if e(k)[p ] @xmath79 and e(l)@xmath80\ne$ ] e(k)@xmath80 $ ] then e(l)[p ] @xmath78 . 5 .
if e(k)[p ] @xmath79 and e(l)[p ] @xmath78 then k does not contain a primitive pth root of unity .
[ p to pp ] 6 . if e(k)[p ] @xmath78 then e(l)@xmath80= $ ] e(k)@xmath80 $ ]
. parts ( 1 ) and ( 2 ) are easily verified .
\3 ) suppose @xmath81 $ ] is trivial . by [ oddpartsadd ]
, it follows that @xmath82 \approx c_p \oplus c_p$ ] so by @xcite*cor 8.1.1 we conclude @xmath4 contains a primitive @xmath64th root of unity .
\4 ) let @xmath83 the largest positive integer such that there is an element of order @xmath84 in @xmath85 .
we have @xmath86= e(k)[p^{m } ] \oplus e^d(k)[p^{m}]$ ] by [ oddpartsadd ] . if @xmath86\ne e(k)[p^{m}]$ ] then @xmath87 \not \approx c_1 $ ] so @xmath82 $ ] @xmath88 . hence @xmath82 \approx c_p$ ] or @xmath89 by @xcite*cor .
6.4 . in the latter case
this would yield @xmath90 \approx c_p \oplus c_p \oplus c_p$ ] which contradicts @xcite*cor .
6.4 , so @xmath90 \approx c_p \oplus c_p $ ] .
\5 ) let @xmath91 be a primtive @xmath64th root of unity .
suppose @xmath92 , and @xmath90 \approx c_p \oplus c_p$ ] .
let @xmath93 gal(@xmath13 ) be nontrivial .
we can choose a basis for @xmath90 $ ] such that the induced galois representation satisfies @xmath94 @xmath95 for some @xmath96 , @xmath97 .
if @xmath98 then @xmath99 so @xmath100 mod @xmath64 . as @xmath101 , @xmath102 so @xmath103 .
as @xmath64 is odd , we conclude @xmath104 , so @xmath105 is the identity .
this means @xmath106 acts trivially on the p - torsion , so @xmath81 \approx
c_p \oplus c_p$ ] , contradicting our hypothesis .
\6 ) suppose @xmath81 \approx c_p \oplus c_p$ ] .
by [ oddpartsadd ] , if @xmath107 \ne e(k)[p^{\infty}]$ ] then @xmath108 \not \approx c_1 $ ] so @xmath82 $ ] @xmath88 .
hence @xmath109 $ ] , contradicting @xcite*cor .
note that there are growths which occur over @xmath110 but not over some quadratic extension : @xmath111 to @xmath112 occurs over @xmath110 but by [ gt ] part 5 not over @xmath113 . on the other hand , there are growths which occur over a quadratic field but not over @xmath110 : @xmath114 to @xmath112 can not over @xmath110 because if it did , @xmath110 would contain a primitive 3rd root of unity by [ gt ] part 3 . on the other hand , @xmath114 to @xmath112 occurs over @xmath113 also , @xmath114 to @xmath115 can not occur over @xmath110 by [ oddpartsadd ] since @xmath116 has no noncuspidal @xmath110-rational points . on the other hand , this growth does occur over @xmath117 : by @xcite*thm 2 there is an elliptic curve e / k with @xmath118 .
choose d @xmath119 , @xmath120 , such that @xmath121 .
then @xmath122 by [ oddpartsadd ] and [ gt ] part 1 .
to study torsion over quadratic fields in the case @xmath123 we use the technique from @xcite . [ count ]
let p be a prime and @xmath124 an elliptic curve with model @xmath125 . 1 .
if a=0 ( i.e. j(e)= 0 ) and p @xmath126 2 mod 3 then @xmath127 and @xmath128 .
if b=0 ( i.e. j(e)=1728 ) and p @xmath126 3 mod 4 then @xmath127 and @xmath128 . if @xmath129 and @xmath130 mod @xmath131 then @xmath127 by @xcite*prop 4.33 . if b=0 and @xmath132 mod @xmath133 then @xmath127 by @xcite*thm 4.23 . now in either case above , @xmath134 by @xcite*thm 4.12 , where @xmath135 and @xmath136 are roots of @xmath137 .
hence @xmath138 [ j=0 ] [ j=1728 ] let @xmath4 be a quadratic field and @xmath7 an elliptic curve . if @xmath139 and @xmath140 is a prime then @xmath38 has no element of order @xmath64 .
if @xmath141 and @xmath142 is a prime then @xmath38 has no element of order @xmath64 .
suppose @xmath139 .
twisting by a square in @xmath143 if necessary , we may assume @xmath32 has a model of the form @xmath144 with @xmath145 .
note that since @xmath143 is a dedekind domain , the principal ideal ( disc(@xmath32 ) ) has only a finite number of prime ideal divisors , and hence disc(@xmath32 ) lies in only a finite number of prime ideals of @xmath143 .
let @xmath146 be a prime in @xmath147 .
since @xmath148 @xmath149 , by the chinese remainder theorem there exists an integer @xmath3 satisfying:@xmath150 @xmath151 furthermore , @xmath152 satisfies the congruences above for every integer @xmath153 , and @xmath154 by the congruences above .
hence by dirichlet s theorem on arithmetic progressions , there are infinitely many primes in this arithmetic progression . in particular , there is a prime @xmath64 satisfying the congruences above such that @xmath32 has good reduction modulo a prime ideal @xmath136 above @xmath64 . as @xmath155=2 $ ]
, we have @xmath156 or @xmath157 . by the comments following @xcite*prop .
3.1 we have an injection of the group @xmath158 $ ] into @xmath159 or @xmath160 .
but by [ count ] we have : @xmath161 @xmath162 as q @xmath163 .
hence in either case ( noting @xmath164 ) , we conclude there is no point of order @xmath148 in @xmath38 .
now suppose @xmath165 .
if @xmath148 is an odd prime , then one can argue just as in the @xmath166 case that there is no point of order @xmath148 .
[ [ section ] ] magma describes the @xmath3-cycle @xmath10 corresponding to an @xmath3-isogeny by providing a polynomial @xmath167 whose roots are precisely the @xmath168-coordinates of the points in @xmath10 . given an @xmath3-isogeny with @xmath3-cycle @xmath10 ,
let @xmath169 denote the field of definition of @xmath10 ( that is , the field obtained by adjoining to @xmath4 all the coordinates of the points of @xmath10 ) and for a polynomial @xmath170 , let @xmath171 denote the splitting field of @xmath170 over @xmath4 .
if an @xmath3-isogeny with @xmath3-cycle @xmath10 is pointwise rational over a field @xmath14 then @xmath167 should split completely over @xmath14 . in particular
if @xmath14 is a quadratic extension of @xmath4 , then @xmath167 must have irreducible factors of degree at most @xmath53 over @xmath4 .
we will now argue that no elliptic curve over @xmath18 has a @xmath172-isogeny pointwise rational over a quadratic extension of @xmath4 .
magma tells us the modular curve @xmath173 has rank @xmath174 , torsion @xmath175 over @xmath4 and @xmath133 cusps over @xmath4 .
the @xmath176 non - cuspidal points correspond to isomorphism classes @xmath177 and using magma we found representatives of each class ( see ) .
as one can see from the table , for each representative @xmath21 with @xmath178 , @xmath167 has an irreducible factor of degree at least @xmath131 and hence @xmath179 \geq [ k(f_c):k ] \geq 3 $ ] .
in particular , there is no quadratic extension @xmath13 such that all the points of @xmath10 are @xmath14-rational .
now since in each of the cases just mentioned , @xmath180 , by @xcite*p .
45 the isomorphism class of @xmath21 just consists of @xmath181 for nonzero @xmath182 , where @xmath183 denotes the image of @xmath10 under quadratic twist by @xmath184 . if @xmath185 then @xmath186 . as @xmath182 , @xmath187 .
hence the @xmath68 isomorphism classes with @xmath178 in do not contain an example of an elliptic curve @xmath7 with a @xmath172-isogeny pointwise rational over a quadratic extension of @xmath4 . in the @xmath188 case ,
magma is not yet able to describe the isomorphism class , so we instead argue as follows : if there is an elliptic curve @xmath7 with a @xmath172-isogeny pointwise rational over a quadratic extension @xmath14 of @xmath4 , then by [ oddpartsadd ] there is an elliptic curve @xmath7 with a point of order 7 over @xmath4 . but
this is impossible by [ j=0 ] .
.(@xmath18 ) representatives ( e , c ) of isomorphism classes corresponding to non - cuspidal k - rational points on a model of @xmath173 [ cols="<,<,<,<",options="header " , ] for each of the six extra points over @xmath0 the grbner basis contains a polynomial @xmath189 . using magma one can check that in each case the only root of @xmath189 over @xmath4 is 0 .
hence the ( @xmath4-rational ) inverse image of these points under @xmath170 is empty . finally , using a grbner basis for the ideal @xmath190 we can determine the non - regular points of @xmath170 . if @xmath191 then the groebner basis contains @xmath192
this has no roots over @xmath4 .
hence the only @xmath4-rational points on @xmath193 are @xmath194 $ ] , @xmath195 $ ] and @xmath196 $ ]
. these points are all cusps so there are no 35-isogenies defined over @xmath197 .
we summarize our findings in the following theorem .
[ isogenies ] if k @xmath73 and @xmath7 is an elliptic curve , @xmath32 has no @xmath198-isogenies pointwise rational over a quadratic extension of @xmath4 for @xmath199 .
if @xmath200 and @xmath7 is an elliptic curve , @xmath32 has no @xmath198-isogenies pointwise rational over a quadratic extension of @xmath4 for @xmath201 .
the curve @xmath173 is genus 1 and rank 1 over @xmath4 . in either case above , there are exactly four elliptic curves over @xmath4 ( up to isomorphism over @xmath4 ) with a 15-isogeny pointwise rational over a quadratic extension of @xmath4 .
[ oddtwist ] let k @xmath202 ) , d @xmath43 @xmath143 , d a nonsquare , and e / k an elliptic curve .
hence the torsion structures @xmath211 and @xmath112 do not grow in any quadratic extension of k. the torsion structures @xmath111 and @xmath212 grow in at most 1 extension , and @xmath114 grows in at most 2 extensions .
let @xmath182 be a non - square .
note that if @xmath213 is a quadratic twist of @xmath32 then @xmath32 is a quadratic twist of @xmath213 ( up to isomorphism over @xmath4 ) . also by [ gt ] ,
all quadratic twists of a curve with odd order torsion will be odd order . by [ najman ] ,
the only odd - order torsion structures occurring over @xmath4 are @xmath214 and @xmath112 .
now if @xmath215 \not = c_1 $ ] and @xmath216 \not = c_1 $ ] then by [ oddpartsadd ] , @xmath217 = c_3 \oplus c_3 $ ] , @xmath218 or @xmath219 , contradicting [ gt ] part 5 or @xcite*cor .
6.4 respectively .
if @xmath220 or @xmath221 , @xmath222 and @xmath223 , then by [ twistback ] , @xmath224 e(l ) , contradicting [ full ] .
if @xmath215 \not = c_1 $ ] and @xmath225 , then by [ twistback ] @xmath226 has a 21-isogeny pointwise rational over a quadratic extension of @xmath4 . but
no such isogeny exists by [ isogenies ] .
if @xmath227 = c_5 $ ] and @xmath225 , then by [ twistback ] @xmath226 has a 35-isogeny pointwise rational over a quadratic extension .
but no such isogeny exists by [ isogenies ] .
if @xmath215 \not = c_1 $ ] and @xmath228 , then by [ twistback ] @xmath226 has a 15-isogeny pointwise rational over a quadratic extension of @xmath4 .
there are four elliptic curves ( two pairs of quadratic twists ) over @xmath4 ( up to isomorphism over @xmath4 ) with such an isogeny . for each such curve @xmath32
( we actually need only check one member of each pair ) , the factorization of the 3-division polynomial of @xmath32 indicates that the nontrivial torsion structures occurring among the quadratic twists of e are @xmath111 and @xmath212 and each occurs exactly once . | let @xmath0 or @xmath1 and let @xmath2 denote the cyclic group of order @xmath3 .
we study how the torsion part of an elliptic curve over @xmath4 grows in a quadratic extension of @xmath4 . in the case @xmath5
\approx c_1 $ ] we investigate how a given torsion structure can grow in a quadratic extension and the maximum number of extensions in which it grows .
we also study the torsion structures which occur as the quadratic twist of a given torsion structure . in order to achieve this we examine @xmath3-isogenies defined over @xmath4 for @xmath6 . |
spontaneous breaking of the chiral symmetry is considered to be the origin of hadron mass .
the chiral symmetry is expected to be ( partially ) restored in finite density and the hadron mass is predicted to decrease , even at the normal nuclear density .
our purpose is to investigate the origin of hadoron mass through mass modification of hadrons .
the dilepton decay channel of vector mesons produced in nuclear reactions is a good probe of in - medium mass modification since it is free from the final state interactions .
we take @xmath0 reaction as an example to explain the expected invariant mass distribution of the vector meson .
a @xmath1 meson produced inside a target nucleus travels and then decays inside or outside the target nucleus .
when the @xmath1 meson decays outside the target nucleus , the mass spectrum is well - known invariant mass in vacuum as in fig .
[ fig : inv](a ) . when the @xmath1 meson decays inside the target nucleus , the observed mass is the one in medium .
so if in - medium mass modification exists , the mass distribution is modified to some extent as in fig .
[ fig : inv](b ) .
what we experimentally measure is the sum of these cases as in fig .
[ fig : inv](c ) .
meson ( a ) in vacuum , and ( b ) in medium .
( c ) is the sum of ( a)+(b ) .
the dashed line indicates the @xmath1 meson mass in vacuum . ]
the kek e325 experiment was performed at kek proton synchrotron to search for in - medium mass modification using the method explained above .
they measured the invariant mass spectra of @xmath2 pairs produced in 12 gev proton beam induced nuclear reactions .
as the nuclear targets , c and cu were used .
the mass resolution was about 11 mev@xmath3 .
figure [ fig : invphie325 ] shows the invariant mass spectrum of @xmath1 meson produced in cu target with @xmath4 @xcite .
the blue line represents an expected line shape assuming mass in vacuum including experimental effects .
there is an excess on the lower side of the @xmath1 mass peak over the expected line shape .
figure [ fig : bg - excess - e325 ] shows the amount of excess versus @xmath5 of @xmath1 mesons .
this figure support the picture that slower @xmath1 meson in larger nuclear target have higher probability to decay inside the nuclear target so that it experience medium modification . to quantitatively extract information on the medium effect , they assume linear dependence of the mass on density as , @xmath6 where @xmath7 is the mass at density @xmath8 , @xmath9 is the normal nuclear density , and @xmath10 is the parameter to be determined .
similarly for the width , they assume @xmath11 where @xmath12 is the width at density @xmath8 , and @xmath13 is the parameter to be determined .
they obtained @xmath14 and @xmath15 , which means that the mass of @xmath1 decreases by 3.4% and the width gets wider by 3.6 times at the normal nuclear density .
they also observed modification of @xmath8 and @xmath16 mass and concluded that @xmath17 , assuming that the parameter is common for @xmath8 and @xmath16 @xcite .
width broadening was not necessary to reproduce the observed invariant mass .
the mass shift parameters @xmath10 for @xmath18 , and @xmath1 are at the same level as the calculations based on qcd sum rule @xcite of @xmath1 measured by kek e325 experiment . ] of @xmath1 measured by kek e325 experiment . ]
the clas g7 experiment at jefferson laboratory used @xmath19 reactions and the light vector mesons were reconstructed using @xmath2 decay channel @xcite .
the results obtained with @xmath20h , c , fe , ti targets were presented . for @xmath16 and @xmath1 meson ,
no mass shift was assumed in the analysis due to their long life .
the @xmath16 and @xmath1 contributions were subtracted to extract the invariant mass spectra of @xmath8 meson .
the mass of @xmath8 meson in nuclear medium does not show any mass shift .
the width are broadened and it is consistent with an expectation from collisional broadening .
dilepton invariant mass spectra were also measured in heavy ion collisions .
ceres /na45 reported @xmath2 invariant mass measured in 158 agev pb+au collisions @xcite .
an in - medium broadening of the @xmath8 mass scenario is favored over a @xmath8 mass dropping scenario .
phenix also reported invariant mass spectra of @xmath2 in au+au collisions at @xmath21a gev @xcite .
an enhancement is observed in the low mass region ( below @xmath1 peak ) .
the enhancement at quite low mass ( @xmath22 gev@xmath3 ) and high @xmath23 ( @xmath24 gev/@xmath25 ) is interpreted as the production of virtual direct photons , which leads to their temperature measurement .
no theoretical models could explain quantitatively the enhancement at the low mass and low @xmath23 region .
there exists some modification in @xmath2 mass spectrum but the origin is not yet clear .
there are even contradiction in the interpretation .
we propose to pursue this problem using the same reaction as kek - e325 but with 100 times more statistics ( @xmath26 ) and with two times better mass resolution ( 11 mev/@xmath27 @xmath28 5 mev/@xmath27 ) .
the proposal was approved as stage-1 and the experiment was named e16 @xcite .
we use 30 gev @xmath29 reactions and measure dilepton invariant mass spectra . as the nuclear target , ch@xmath30 , c , cu , pb
the j - parc e16 experiment has the following advantages and disadvantages compared to other experiments .
it observes @xmath2 decay channel so it can eliminate final state interactions in contrast to the case of experiments using hadronic decay channels .
however , @xmath2 decay channel has very tiny branching ratio ( @xmath31 for @xmath1 meson ) .
the e16 experiment uses proton induced reactions , therefore , the system is cold and static so is simpler compared to that of heavy ion collisions .
the e16 experiment is expected to measure @xmath1 meson invariant mass modification .
compared to @xmath8 and @xmath16 mesons , @xmath1 meson has a non - overlapping separated peak and a narrower width .
however , the production cross section of @xmath1 meson is much smaller than that of @xmath8 and @xmath16 , thus it is difficult to collect high statistics and clas - g7 and ceres can not discuss the mass spectra of @xmath1 .
the disadvantages , mainly come from the fact that @xmath32 is a rare probe , are overcome by collecting high statistics data . when the statistics is achieved , the invariant mass distribution of slowly moving @xmath1 meson whose @xmath5 is less than 0.5 which is obtained with pb target is expected to have double peak as in fig .
[ fig : invmass - phi - e16 ] .
note that the modification parameters obtained by kek - e325 are assumed .
the @xmath5 and the target size dependence of the modification expected to be obtained is in fig . [
fig : bg - excess - e16 ] .
so more systematic study is possible .
we are able to obtain dispersion relation as the blue points in fig .
[ fig : dispersion ] , which is qualitatively new information .
these new information can give further insight on the in - medium modification .
of @xmath1 expected to be obtained by j - parc e16 experiment . ] of @xmath1 expected to be obtained by j - parc e16 experiment . ] and is extrapolated to 3 gev/@xmath25 .
black dotted curve shows the uncertainty of the calculation .
blue points shows the statistical uncertainties expected to be obtained by the j - parc e16 experiment .
the center values are taken from the theoretical calculation mentioned above .
purple point is the results obtained by kek e325 experiment . ] to achieve 100 times more statistics , we utilize 10 times more intense beam ( @xmath33 protons per pulse ( ppp ) @xmath28 @xmath34 ppp ) , a spectrometer with 5 times larger acceptance , and 2 times larger production rate due to the increased beam energy ( 12 gev @xmath28 30 gev ) .
the j - parc e16 experiment plans to use the high momentum beam line which will be constructed at j - parc hadron experimental facility .
j - parc , japan proton accelerator research complex , is a high intensity proton accelerator and is located at tokai village in japan .
the main ring ( mr ) of j - parc can accelerate protons up to 30 gev .
figure [ fig : sy - hd ] shows the plan view of the switchyard and the hadron experimental facility .
the protons in the mr are slowly extracted to line - a .
the proton beam follows line - a through the switchyard and is delivered to the hadron experimental facility .
the protons collide the t1 target to provide secondary beams to the existing beam lines such as k1.8 , k1.8br , and kl .
the beam power was 24 kw as of 2013 , which corresponds to @xmath35 ppp . to make primary proton beam available to e16 experiment , the high momentum beam line which is called line - b
is being constructed .
line - b borrows a small piece of the beam ( @xmath36 ) in line - a with a lambertson - type magnet at the switchyard .
the beam is extracted to the south side of the hadron experimental facility where e16 spectrometer is to be built .
a 3d view of the j - parc e16 spectrometer is shown on the left side of fig .
[ fig : spectrometer ] .
the e16 detectors are all installed inside a giant dipole magnet with a field strength of 1.7 t at the center .
a horizontal cut view at the center is presented on the right side of fig .
[ fig : spectrometer ] .
the proton beam runs from bottom to the top of the figure , and hit the target at the center of the spectrometer .
the spectrometer consists of gem trackers ( gtr ) @xcite , hadron blind detectors ( hbd ) @xcite , and lead glass ( lg ) calorimeters .
a module is defined as a set of gtr , hbd and lg which covers 30 degrees both horizontally and vertically .
the full design of the spectrometer consists of 26 modules .
gtr is made of three layers of position - sensitive gem tracking chambers with the sizes of 100 @xmath37 100 mm@xmath20 , 200 @xmath37 200 mm@xmath20 and 300 @xmath37 300 mm@xmath20 , respectively .
hbd is a cherenkov counter and is used for electron identification together with lg .
particle tracks in the magnetic field are reconstructed with gtr so that the momenta are measured .
electron candidates are selected with hbd and lg .
position resolution of 100 @xmath38 with incident angles of up to 30 degrees is required for gtr .
rejection factors of 100 and 25 are required for hbd and lg , respectively .
100 mm@xmath20 and 200 @xmath37 200 mm@xmath20 , respectively . ] 100 mm@xmath20 and 200 @xmath37 200 mm@xmath20 , respectively . ]
a schematic of a gem tracking chamber is shown in fig .
[ fig : gem ] .
ar + co2 ( 70:30 ) is used as the amplification gas .
three gem foils are placed and they amplify the ionization electrons produced by a traversing charged particle in the gap above the top gem .
the amplified signal is readout with two dimensional strip readout board .
a custom preamp board using apv25 chip @xcite has been developed .
the mass production type of gem tracking chambers with three different sizes and preamp boards have been built .
the performance of them was evaluated with charged particle beams at j - parc and elph .
the required resolution of 100 @xmath38 was achieved for incident angles of up to 30 degrees . a picture of the gem chambers with two sizes are shown in fig .
[ fig : picgem ] .
first level trigger is readout from the bottom of the gem foil . a prototype of asd ( amplifier - shaper - discriminator ) asic for the trigger readout has also been developed .
hbd is a type of cherenkov detector using csi evaporated gem as a photocathode .
our hbd has been developed based on the phenix hbd experience @xcite .
cf@xmath39 serves as radiator and amplification gas . with the radiator length of 50 cm
, 11 photoelectrons are expected .
a schematic of the photocathode is shown in fig .
[ fig : hbd - schematics ] .
the incident electron emit cherenkov photons .
the photons are converted into photoelectrons by the csi layer which is evaporated on top of the top gem .
the photoelectrons are then amplified by the gems .
a weak reverse bias field is applied in the gap between the mesh and the top gem , so that the ionization electrons in the gap are swept into the mesh .
even with the reverse bias field , photoelectrons produced near the top gem surface are still attracted by the gem s field and are amplified to be detected .
therefore , hbd is blind to ionization while is sensitive to cherenkov photons .
the size of the photocathode of a hbd module is 600 @xmath37 600 mm@xmath20 and four photocathodes with a size of 300 @xmath37 300 mm@xmath20 are used to fill the module .
extensive r & d effort has been performed to establish hbd components such as efficient and robust csi gems , airtight chambers and readout boards .
prototypes of gems with and without csi , chambers and readout boards in small sizes and real sizes were produced .
a picture of a prototype of the hbd chamber in real size is shown in fig .
[ fig : hbdpic ] .
it corresponds to a module of the spectrometer .
a beam test was performed with negatively charged particle beam of 1.0 gev/@xmath25 at the j - parc k1.1br to evaluate the performance of a prototype hbd in small size .
a pion rejection factor of 100 with an electron efficiency of 80% was achieved using cluster size analysis .
the prototype of hbd in real size also operates well .
the performance meet the required rejection and efficiency for the experiment .
the experiment was approved as stage-1 in 2007 .
detector r&d started in 2008 .
the construction budget of the high - p beam line was approved in 2013 .
technical design report was submitted and the mass production of detectors started in 2014 .
due to the budgetary limitation , we start with one third of the full design .
the one third of the full design will be ready for the first physics run which is anticipated in jfy2016 .
j - parc e26 experiment has been proposed to investigate @xmath16 meson in nuclear medium@xcite . it plan to use @xmath40 beam at j - parc k1.8 beam line with a momentum of 1.8 gev/@xmath25 and with an intensity of @xmath41 / pulse . the reaction @xmath42 is used .
invariant mass of @xmath16 meson is measured with @xmath43 decay mode .
when neutron is detected at zero degree , recoilless @xmath16 production is realized .
the condition is suitable for the study of in - medium effect .
nuclear @xmath16 bound state can be searched via forward neutron measurement .
j - parc e29 experiment has been proposed to investigate in - medium mass modification of @xmath1 meson via @xmath1 meson bound state in target nucleus@xcite .
it plan to use @xmath44 beam with a momentum of 1.1 gev/@xmath25 and with an intensity of @xmath45/pulse .
when four strangeness are identified in the final state , the double @xmath1 production , @xmath46 , dominates .
the forward - going @xmath1 meson is detected via @xmath47 decay .
the @xmath1 meson in nucleus is detected via @xmath48 decay , which occur only when @xmath1 is in nucleus ( @xmath49 ) .
missing mass spectrum is calculated with the beam momentum and the forward - going @xmath1 momentum .
the backward @xmath1 is at the same order of fermi momentum which is detected via @xmath48 decay in nucleus .
when high intensity high resolution secondary beam line ( hihr ) which was proposed by rcnp is realized , experimental study of @xmath1 meson in nuclear medium using a similar method as j - parc e26 can be done .
a @xmath50pulse @xmath40 beam with a momentum of @xmath51 gev is used to induce the @xmath52 reaction . if the neutron is identified at the forward angle , ultra slow @xmath1 is selected . forward
neutron measurement may lead to observation of nuclear @xmath1 bound state .
about 10 times more @xmath1 compared to e16 is expected to be collected with @xmath53 .
the origin of hadron mass is studied through mass modification of vector mesons .
there are many measurements of dilepton invariant mass in hot and cold system .
there exists some modification but the origin is not yet clear .
j - parc e16 experiment pursue it by collecting 100 times more statics compare to the kek e325 experiment .
we expect to obtain double peak structure in @xmath1 meson invariant mass spectra , wide range of system size dependence of the in - medium modification , and the dispersion relation of @xmath1 meson in nuclear medium .
we start with one - third of the design configuration and physics run is anticipated in jfy2016 .
more experiments regarding vector meson mass modification are planned and whole together provide further insights on the origin of mass , and the chiral symmetry .
we would like to give our thanks to the staff of kek fuji test beam line , elph at tohoku university , leps at spring-8 , j - parc hadron experimental facility and riken ri beam factory for their support for the beam test of detectors .
we also would like to thank to kek electronics system group ( e - sys ) for their help in the development and test of the readout circuits .
this study was partly supported by grant - in - aid for jsps fellows 12j01196 , riken spdr program , and mext / jsps kakenhi grant numbers 19654036 , 19340075 , 21105004 and 26247048 . `
http://j-parc.jp/researcher/hadron/en/pac_0903/pdf/ozawa.pdf ` ` http://j-parc.jp/researcher/hadron/en/pac_1007/pdf/ ` + `
kek_j - parc - pac2010 - 08.pdf ` ` http://j-parc.jp/researcher/hadron/en/pac_0907/pdf/ ` + ` ohnishi.pdf ` ` http://j-parc.jp/researcher/hadron/en/pac_1007/pdf/ ` + ` kek_j - parc - pac2010 - 02.pdf ` | study of in - medium mass modification has attracted interest in terms of the restoration of the spontaneously broken chiral symmetry , which is responsible for the generation of hadron mass .
many experiments were performed to measure in - medium property of hadrons but there is no consensus yet .
j - parc e16 has been proposed to study in - medium property of vector mesons via dilepton decay channel .
the status of spectrometer r & d is explained .
other related experiments planned at j - parc are also introduced . |
recent years have witnessed much interest in cognitive radio systems due to their promise as a technology that enables systems to utilize the available spectrum much more effectively .
this interest has resulted in a spur of research activity in the area . in @xcite , asghari and
aissa , under constraints on the average interference caused at the licensed user over rayleigh fading channels , studied two adaptation policies at the secondary user s transmitter in a cognitive radio system one of which is variable power and the other is variable rate and power .
they maximized the achievable rates under the above constraints and the bit error rate ( ber ) requirement in mqam modulation .
the authors in @xcite derived the fading channel capacity of a secondary user subject to both average and peak received - power constraints at the primary receiver .
in addition , they obtained optimum power allocation schemes for three different capacity notions , namely , ergodic , outage , and minimum - rate .
_ in @xcite studied the performance of spectrum - sensing radios under channel fading .
they showed that due to uncertainty resulting from fading , local signal processing alone may not be adequate to meet the performance requirements .
therefore , to remedy this uncertainty they also focused on the cooperation among secondary users and the tradeoff between local processing and cooperation in order to maximize the spectrum utilization .
furthermore , the authors in @xcite focused on the problem of designing the sensing duration to maximize the achievable throughput for the secondary network under the constraint that the primary users are sufficiently protected .
they formulated the sensing - throughput tradeoff problem mathematically , and use energy detection sensing scheme to prove that the formulated problem indeed has one optimal sensing time which yields the highest throughput for the secondary network .
moreover , quan _
_ in @xcite introduced a novel wideband spectrum sensing technique , called multiband joint detection , which jointly detects the signal energy levels over multiple frequency bands rather than considering one band at a time . in many wireless systems
, it is very important to provide reliable communications while sustaining a certain level of quality - of - service ( qos ) under time - varying channel conditions .
for instance , in wireless multimedia transmissions , stringent delay qos requirements need to be satisfied in order to provide acceptable performance levels . in cognitive radio systems ,
challenges in providing qos assurances increase due to the fact that secondary users should operate under constraints on the interference levels that they produce to primary users . for the secondary users , these interference constraints lead to variations in transmit power levels and channel accesses .
for instance , intermittent access to the channels due to the activity of primary users make it difficult for the secondary users to satisfy their own qos limitations .
these considerations have led to studies that investigate the cognitive radio performance under qos constraints .
musavian and aissa in @xcite considered variable - rate , variable - power mqam modulation employed under delay qos constraints over spectrum - sharing channels . as a performance metric , they used the effective capacity to characterize the maximum throughput under qos constraints .
they assumed two users sharing the spectrum with one of them having a primary access to the band .
the other , known as secondary user , is constrained by interference limitations imposed by the primary user .
considering two modulation schemes , continuous mqam and discrete mqam with restricted constellations , they obtained the effective capacity of the secondary user s link , and derived the optimum power allocation scheme that maximizes the effective capacity in each case .
additionally , in @xcite , they proposed a qos constrained power and rate allocation scheme for spectrum sharing systems in which the secondary users are allowed to use the spectrum under an interference constraint by which a minimum - rate of transmission is guaranteed to the primary user for a certain percentage of time .
moreover , applying an average interference power constraint which is required to be fulfilled by the secondary user , they obtained the maximum arrival - rate supported by a rayleigh block - fading channel subject to satisfying a given statistical delay qos constraint .
we note that in these studies on the performance under qos limitations , channel sensing is not incorporated into the system model . as a result ,
adaptation of the cognitive transmission according to the presence or absence of the primary users is not considered . in @xcite , where we also concentrated on cognitive transmission under qos constraint
, we assumed that the secondary transmitter sends the data at two different fixed rates and power levels , depending on the activity of the primary users , which is determined by channel sensing performed by the secondary users .
we constructed a state transition model with eight states to model this cognitive transmission channel , and determined the effective capacity . on the other hand
, we assumed in @xcite that channel sensing is done only in one channel , and did not impose explicit interference constraints . in this paper , we study the effective capacity of cognitive radio channels where the cognitive radio detects the activity of primary users in a multiband environment and then performs the data transmission in one of the transmission channels .
both the secondary receiver and the secondary transmitter know the fading coefficients of their own channel , and of the channel between the secondary transmitter and the primary receiver .
the cognitive radio has two power allocation policies depending on the activities of the primary users and the sensing decisions .
more specifically , the contributions of this paper are the following : 1 .
we consider a scenario in which the cognitive system employs multi - channel sensing and uses one channel for data transmission thereby decreasing the probability of interference to the primary users .
we identify a state - transition model for cognitive radio transmission in which we compare the transmission rates with instantaneous channel capacities , and also incorporate the results of channel sensing .
3 . we determine the effective capacity of the cognitive channel under limitations on the average interference power experienced by the primary receiver .
we identify the optimal criterion to select the transmission channel out of the available channels and obtain the optimal power adaptation policies that maximize the effective capacity .
we analyze the interactions between the effective capacity , qos constraints , channel sensing duration , channel detection threshold , detection and false alarm probabilities through numerical techniques .
the organization of the rest of the paper is as follows : in section [ system ] , we discuss the channel model and analyze multi - channel sensing .
we describe the channel state transition model in section [ subsec : state ] under the assumption that the secondary users have perfect csi and send the data at rates equal to the instantaneous channel capacity values . in section [ sec : consraint ] , we analyze the received interference power at the primary receiver and apply this as a power constraint on the secondary users . in section
[ effective capacity ] , we define the effective capacity and find the optimal power distribution and show the criterion to choose the best channel .
numerical results are shown in section [ numericalresults ] , and conclusions are provided in section [ conlusion ] .
in this paper , we consider a cognitive radio system in which secondary users sense @xmath0 channels and choose one channel for data transmission .
we assume that channel sensing and data transmission are conducted in frames of duration @xmath1 seconds . in each frame
, @xmath2 seconds is allocated for channel sensing while data transmission occurs in the remaining @xmath3 seconds .
transmission power and rate levels depend on the primary users activities .
if all of the channels are detected as busy , transmitter selects one channel with a certain criterion , and sets the transmission power and rate to @xmath4 and @xmath5 , respectively , where @xmath6 is the index of the selected channel and @xmath7 denotes the time index . note that if @xmath8 , transmitter stops sending information when it detects primary users in all channels .
if at least one channel is sensed to be idle , data transmission is performed with power @xmath9 and at rate @xmath10 .
if multiple channels are detected as idle , then one idle channel is selected again considering a certain criterion .
the discrete - time channel input - output relation between the secondary transmitter and receiver in the @xmath11 symbol duration in the @xmath12 channel is given by @xmath13 if the primary users are absent .
on the other hand , if primary users are present in the channel , we have @xmath14 where @xmath15 and @xmath16 denote the complex - valued channel input and output , respectively . in ( [ input - out1 ] ) and ( [ input - out2 ] ) , @xmath17 is the channel fading coefficient between the cognitive transmitter and the receiver .
we assume that @xmath17 has a finite variance , i.e. , @xmath18 , but otherwise has an arbitrary distribution .
we define @xmath19 .
we consider a block - fading channel model and assume that the fading coefficients stay constant for a block of duration @xmath1 seconds and change from one block to another independently in each channel . in ( [ input - out2 ] ) , @xmath20 represents the active primary user s faded signal arriving at the secondary receiver in the @xmath21 channel , and has a variance @xmath22 .
@xmath23 models the additive thermal noise at the receiver , and is a zero - mean , circularly symmetric , complex gaussian random variable with variance @xmath24 for all @xmath25 .
we assume that the bandwidth of the @xmath21 channel is @xmath26 . in the absence of detailed information on primary users transmission policies ,
energy - based detection methods are favorable for channel sensing . knowing that wideband channels exhibit frequency selective features
, we can divide the band into channels and estimate each received signal through its discrete fourier transform ( dft ) @xcite .
the channel sensing can be formulated as a hypothesis testing problem between the noise @xmath23 and the signal @xmath20 in noise .
noting that there are @xmath27 complex symbols in a duration of @xmath2 seconds in each channel with bandwidth @xmath28 , the hypothesis test in channel @xmath29 can mathematically be expressed as follows : @xmath30 for the above detection problem , the optimal neyman - pearson detector is given by @xcite @xmath31 we assume that @xmath20 has a circularly symmetric complex gaussian distribution with zero - mean and variance @xmath32 . assuming further that @xmath33 are i.i.d . , we can immediately conclude that the test statistic @xmath34 is chi - square distributed with @xmath35 degrees of freedom . in this case , the probabilities of false alarm and detection can be established as follows : @xmath36 where @xmath37 denotes the regularized lower gamma function and is defined as @xmath38 where @xmath39 is the lower incomplete gamma function and @xmath40 is the gamma function . in figure
[ fig : fig1 ] , the probability of detection , @xmath41 , and the probability of false alarm , @xmath42 , are plotted as a function of the energy detection threshold , @xmath43 , for different values of channel detection duration .
note that the bandwidth is @xmath44khz and the block duration is @xmath45s .
we can see that when the detection threshold is low , @xmath41 and @xmath42 tend to be 1 , which means that the secondary user , always assuming the existence of an active primary user , transmits with power @xmath46 and rate @xmath47 . on the other hand ,
when the detection threshold is high , @xmath41 and @xmath42 are close to zero , which means that the secondary user , being unable to detect the activity of the primary users , always transmits with power @xmath48 and rate @xmath49 , possibly causing significant interference .
the main purpose is to keep @xmath41 as close to 1 as possible and @xmath42 as close to 0 as possible .
therefore , we have to keep the detection threshold in a reasonable interval .
note that the duration of detection is also important since increasing the number of channel samples used for sensing improves the quality of channel detection . in the hypothesis testing problem in ( [ hypothesis ] ) , another approach is to consider @xmath34 as gaussian distributed , which is accurate if @xmath27 is large @xcite . in this case , the detection and false alarm probabilities can be expressed in terms of gaussian @xmath50-functions .
we would like to note the rest of the analysis in the paper does not depend on the specific expressions of the false alarm and detection probabilities .
however , numerical results are obtained using ( [ false alarm ] ) and ( [ eq : probdetect ] ) .
in this paper , we assume that both the secondary receiver and transmitter have perfect channel side information ( csi ) , and hence perfectly know the realizations of the fading coefficients @xmath51 .
we further assume that the wideband channel is divided into channels , each with bandwidth that is equal to the coherence bandwidth @xmath52 .
therefore , we henceforth have @xmath53 . with this assumption
, we can suppose that independent flat fading is experienced in each channel . in order to further simplify the setting
, we consider a symmetric model in which fading coefficients are identically distributed in different channels .
moreover , we assume that the background noise and primary users signals are also identically distributed in different channels and hence their variances @xmath54 and @xmath55 do not depend on @xmath29 , and the prior probabilities of each channel being occupied by the primary users are the same and equal to @xmath56 . in channel
sensing , the same energy threshold , @xmath43 , is applied in each channel .
finally , in this symmetric model , the transmission power and rate policies when the channels are idle or busy are the same for each channel . due to the consideration of a symmetric model , we in the subsequent analysis drop the subscript @xmath29 in the expressions for the sake of brevity .
first , note that we have the following four possible scenarios considering the correct detections and errors in channel sensing : _ scenario 1 : _ all channels are detected as busy , and channel used for transmission is actually busy .
_ scenario 2 : _ all channels are detected as busy , and channel used for transmission is actually idle . _
scenario 3 : _ at least one channel is detected as idle , and channel used for transmission is actually busy . _
scenario 4 : _ at least one channel is detected as idle , and channel used for transmission is actually idle . in each scenario
, we have one state , namely either on or off , depending on whether or not the instantaneous transmission rate exceeds the instantaneous channel capacity . considering the interference @xmath57 caused by the primary users as additional gaussian noise , we can express the instantaneous channel capacities in the above four scenarios as follows : [ channel capacity ] _ scenario 1 : _ @xmath58 . _
scenario 2 : _ @xmath59 .
_ scenario 3 : _ @xmath60 .
_ scenario 4 : _ @xmath61 .
above , we have defined @xmath62 note that @xmath63 denotes the fading power . in scenarios 1 and 2 ,
the secondary transmitter detects all channels as busy and transmits the information at rate @xmath64 on the other hand , in scenarios 3 and 4 , at least one channel is sensed as idle and the transmission rate is @xmath65 since the transmitter , assuming the channel as idle , sets the power level to @xmath66 and expects that no interference from the primary transmissions will be experienced at the secondary receiver ( as seen by the absence of @xmath55 in the denominator of @xmath67 ) . in scenarios 1 and 2 ,
transmission rate is less than or equal to the instantaneous channel capacity .
hence , reliable transmission at rate @xmath47 is attained and channel is in the on state .
similarly , the channel is in the on state in scenario 4 in which the transmission rate is @xmath49 . on the other hand , in scenario 3
, transmission rate exceeds the instantaneous channel capacity ( i.e. , @xmath68 ) due to miss - detection . in this case
, reliable communication can not be established , and the channel is assumed to be in the off state .
note that the effective transmission rate in this state is zero , and therefore information needs to be retransmitted .
we assume that this is accomplished through a simple arq mechanism . for this cognitive transmission model , we initially construct a state transition model .
while the ensuing discussion describes this model , figure [ fig : fig2 ] provides a depiction . as seen in fig .
[ fig : fig2 ] , there are @xmath69 on states and 1 off state .
the single off state is the one experienced in scenario 3 .
the first on state , which is the top leftmost state in fig .
[ fig : fig2 ] , is a combined version of the on states in scenarios 1 and 2 in both of which the transmission rate is @xmath47 and the transmission power is @xmath46 .
note that all the channels are detected as busy in this first on state .
the remaining on states labeled @xmath70 through @xmath71 can be seen as the expansion of the on state in scenario 4 in which at least one channel is detected as idle and the channel chosen for transmission is actually idle .
more specifically , the @xmath12 on state for @xmath72 is the on state in which @xmath73 channels are detected as idle and the channel chosen for transmission is idle .
note that the transmission rate is @xmath49 and the transmission power is @xmath48 in all on states labeled @xmath70 through @xmath71 .
next , we characterize the state transition probabilities .
state transitions occur every @xmath1 seconds .
we can easily see that the probability of staying in the first on state , in which all channels are detected as busy , is expressed as follows : @xmath74 where @xmath75 is the probability that channel is detected as busy , and @xmath41 and @xmath42 are the probabilities of detection and false alarm , respectively as defined in ( [ eq : probdetect ] ) . recall that @xmath56 denotes the probability that a channel is busy ( i.e. , there are active primary users in the channel ) .
it is important to note that the transition probability in ( [ eq : transitionprob ] ) is obtained under the assumptions that the primary user activity is independent among the channels and also from one block to another . indeed , under the assumption of independence over the blocks , the state transition probabilities do not depend on the originating state and hence we have @xmath76 where we have defined @xmath77 for all @xmath78 .
similarly , we can obtain for @xmath79 , @xmath80 now , we can easily observe that the transition probabilities for the off state are @xmath81 then , we can easily see that the @xmath82 state transition probability matrix can be expressed as @xmath83 = \left [ \begin{array}{cccc } p_{1 } & . & .
& p_{m+2 } \\ .&\quad&\quad&.\\ .&\quad&\quad&.\\
p_{1 } & . & . &
p_{m+2}\\ \end{array } \right]$ ] note that @xmath84 has a rank of 1 .
note also that in each frame duration of @xmath1 seconds , @xmath85 bits are transmitted and received in state 1 , and @xmath86 bits are transmitted and received in states 2 through @xmath69 , while the transmitted number of bits is assumed to be zero in state @xmath87 .
in this section , we consider interference power constraints to limit the transmission powers of the secondary users and provide protection to primary users . in particular , we assume that the transmission power of the secondary users is constrained in such a way that the average interference power on the primary receiver is limited .
note that interference to the primary users is caused in scenarios 1 and 3 . in scenario 1 ,
the channel is busy , and the secondary user , detecting the channel as busy , transmits at power level @xmath88 .
consequently , the instantaneous interference power experienced by the primary user is @xmath89 where @xmath90 is the magnitude - square of the fading coefficient of the channel between the secondary transmitter and the primary user .
note also that the probability of being in scenario 1 ( i.e. , the probability of detecting all channels busy and having the chosen transmission channel as actually busy ) is @xmath91 , as can be easily seen through an analysis similar to that in ( [ eq : transitionprobformula ] ) . in scenario 3 ,
the secondary user , detecting the channel as idle , transmits at power @xmath92 although the channel is actually is busy . in this case
, the instantaneous interference power is @xmath93 . since we consider power adaption , transmission power levels @xmath88 and @xmath92 in general vary with @xmath94 and also with @xmath95 , which is the power of the fading coefficient between the secondary transmitter and secondary receiver in the chosen transmission channel . hence ,
in both scenarios , the instantaneous interference power levels depend on both @xmath94 and @xmath95 whose distributions depend on the criterion with which the transmission channel is chosen and the number of available channels from which the selection is performed . for this reason ,
it is necessary in scenario 3 to separately consider the individual cases with different number of idle - detected channels .
we have @xmath0 such cases .
for instance , in the @xmath21 case for @xmath96 , we have @xmath29 channels detected as idle and the channel chosen out of these @xmath29 channels is actually busy .
the probability of the @xmath21 case can be easily found to be @xmath97 following the above discussion , we can now express the average interference constraints as follows : @xmath98 note from above that @xmath99 is the constraint on the interference averaged over the distributions of @xmath95 and @xmath94 ( through the expectations ) , and also averaged over the probabilities of different scenarios and cases .
it is important to note that the term @xmath100 , as discussed above , depends in general on the number of idle - detected channels , @xmath29 .
this dependence is indicated through the subscript @xmath29 . in a system with more strict requirements on the interference ,
the following individual interference constraints can be imposed : @xmath101 if , for instance , @xmath102 , then interference averaged over fading is limited by the same constraint regardless of which scenario is being realized . as considered in @xcite , by appropriately choosing the values of @xmath103 and @xmath104 in ( [ eq : individualinterference ] ) , we can provide primary users a minimum rate guarantee for a certain percentage of the time in a rayleigh fading environment through the following outage constraints : @xmath105 @xmath106 and @xmath107 can be seen as the outage constraints in scenario 1 and in the @xmath21 case of scenario 3 , respectively . in the above formulations , @xmath108 is the required minimum transmission rate to be provided to the primary users with outage probabilities @xmath106 and @xmath107 , and @xmath109 where @xmath110 is the fading coefficient of the channel between the primary transmitter and primary receiver .
@xmath111 is the variance of the zero - mean , circularly symmetric , complex gaussian thermal noise at the primary receiver .
@xmath112 is the transmission power of the primary transmitter . under the assumption that @xmath113 is an exponential random variable ( i.e.
, we have a rayleigh fading channel between the primary transmitter and receiver ) , the outage probability in ( [ outage1 ] ) can be expressed as follows : @xmath114 where ( [ eq : outageconstraint2 ] ) is obtained by performing integration with respect to the probability density function ( pdf ) of @xmath113 in the evaluation of the probability expression in ( [ eq : outageconstraint1 ] ) . as a result ,
the expectation in ( [ eq : outageconstraint2 ] ) is with respect to the remaining random components @xmath88 and @xmath94 .
finally , the inequality in ( [ eq : outageconstraint3 ] ) follows from the concavity of the function @xmath115 and jensen s inequality . from ( [ eq : outageconstraint3 ] ) , we can immediately see that if we impose @xmath116 then the constraint in ( [ outage1 ] ) will be satisfied .
a similar discussion follows for ( [ outage2 ] ) as well . in the subsequent parts of the paper , we assume that an average interference power constraint in the form given in ( [ average interference power ] ) is imposed .
in this section , we identify the maximum throughput that the cognitive radio channel with the aforementioned state - transition model can sustain under interference power constraints and statistical qos limitations imposed in the form of buffer or delay violation probabilities . wu and negi in @xcite defined the effective capacity as the maximum constant arrival rate that can be supported by a given channel service process while also satisfying a statistical qos requirement specified by the qos exponent @xmath117 .
if we define @xmath50 as the stationary queue length , then @xmath117 is defined as the decay rate of the tail distribution of the queue length @xmath50 : @xmath118 hence , we have the following approximation for the buffer violation probability for large @xmath119 : @xmath120 .
therefore , larger @xmath117 corresponds to more strict qos constraints , while the smaller @xmath117 implies looser constraints . in certain settings , constraints on the queue length
can be linked to limitations on the delay and hence delay - qos constraints .
it is shown in @xcite that @xmath121 for constant arrival rates , where @xmath122 denotes the steady - state delay experienced in the buffer . in the above formulation
, @xmath123 is a positive constant , @xmath124 and @xmath125 is the source arrival rate .
therefore , effective capacity provides the maximum arrival rate when the system is subject to statistical queue length or delay constraints in the forms of @xmath126 or @xmath127 , respectively .
since the average arrival rate is equal to the average departure rate when the queue is in steady - state @xcite , effective capacity can also be seen as the maximum throughput in the presence of such constraints .
the effective capacity for a given qos exponent @xmath117 is given by @xmath128 where @xmath129 is the time - accumulated service process , and @xmath130 is defined as the discrete - time , stationary and ergodic stochastic service process .
note that @xmath131 is the asymptotic log - moment generating function of @xmath132 , and is given by @xmath133.\ ] ] the service rate according to the model described in section [ subsec : state ] is @xmath134 if the cognitive system is in state 1 at time @xmath29 .
similarly , the service rate is @xmath135 in the states between 2 and @xmath69 . in the off state ,
instantaneous transmission rate exceeds the instantaneous channel capacity and reliable communication can not be achieved .
therefore , the service rate in this state is effectively zero . in the next result
, we provide the effective capacity for the cognitive radio channel and state transition model described in the previous section .
[ theo : effectcap ] for the cognitive radio channel with the state transition model given in section [ subsec : state ] , the normalized effective capacity ( in bits / s / hz ) under the average interference power constraint ( [ average interference power ] ) is given by @xmath136 above , @xmath137 for @xmath138 denote the state transition probabilities defined in ( [ eq : transitionprob1 ] ) , ( [ eq : transitionprobk ] ) , and ( [ eq : transitionprobm+2 ] ) in section [ subsec : state ] .
note also that the maximization is with respect to the power adaptation policies @xmath88 and @xmath92 .
_ remark : _ in the effective capacity expression ( [ eq : effectivecap ] ) , the expectation @xmath139 in the constraint and @xmath140 are with respect to the joint distribution of @xmath141 of the channel selected for transmission when all channels are detected busy .
the expectations @xmath142 and @xmath143 are with respect to the joint distribution of @xmath141 of the channel selected for transmission when @xmath29 channels are detected as idle .
_ proof of theorem [ theo : effectcap ] : _ in ( * ? ? ? * chap .
7 , example 7.2.7 ) , it is shown for markov modulated processes that @xmath144 where @xmath145 is the spectral radius ( i.e. , the maximum of the absolute values of the eigenvalues ) of the matrix @xmath146 , @xmath84 is the transition matrix of the underlying markov process , and @xmath147 is a diagonal matrix whose components are the moment generating functions of the processes in given states .
the rates supported by the cognitive radio channel with the state transition model described in the previous section can be seen as a markov modulated process and hence the setup considered in @xcite can be immediately applied to our setting . since the processes in the states are time - varying transmission rates , we can easily find that @xmath148 .
then , we have @xmath149 $ ] .
since @xmath146 is a matrix with unit rank , we can readily find that @xmath150 then , combining ( [ eq : sp ] ) with ( [ eq : theta - envelope ] ) and ( [ exponent ] ) , normalizing the expression with @xmath151 in order to have the effective capacity in the units of bits / s / hz , and considering the maximization over power adaptation policies , we reach to the effective capacity formula given in ( [ eq : effectivecap ] ) .
@xmath152 we would like to note that the effective capacity expression in ( [ eq : effectivecap ] ) is obtained for a given sensing duration @xmath2 , detection threshold @xmath43 , and qos exponent @xmath117 . in the next section , we investigate the impact of these parameters on the effective capacity through numerical analysis . before the numerical analysis
, we first identify below the optimal power adaptation policies that the secondary users should employ .
the optimal power adaptations for the secondary users under the constraint given in ( [ average interference power ] ) are @xmath153 , & \hbox{$\frac{z}{z_{sp}}\geq\beta_{1}\lambda$ } \\ 0 , & \hbox{otherwise } \end{array } \right.,\ ] ] and @xmath154 , & \hbox{$\frac{z}{z_{sp}}\geq\beta_{2}\lambda$ } \\ 0 , & \hbox{otherwise }
\end{array } \right.,\ ] ] where @xmath155 , @xmath156 , @xmath157 , @xmath158 and @xmath159 .
@xmath160 is a parameter whose value can be found numerically by satisfying the constraint ( [ average interference power ] ) with equality .
_ proof : _ since logarithm is a monotonic function , the optimal power adaptation policies can also be obtained from the following minimization problem : @xmath161 it is clear that the objective function in ( [ objectivefunc ] ) is strictly convex and the constraint function in ( [ average interference power ] ) is linear with respect to @xmath162 and @xmath163 . and
@xmath164 in ( [ rate1 ] ) and ( [ rate2 ] ) with respect to @xmath88 and @xmath92 respectively , strict convexity of the exponential function , and the fact that the nonnegative weighted sum of strictly convex functions is strictly convex ( * ? ? ?
* section 3.2.1 ) . ] then , forming the lagrangian function and setting the derivatives of the lagrangian with respect to @xmath162 and @xmath163 equal to zero , we obtain : @xmath165\alpha^{m}f_{}(z , z_{sp})=0\label{lamda1}\\ & \hspace{-.5cm}\left[\lambda\rho(1-p_{d})z_{sp}-\frac{c(1-\rho)(1-p_{f})z}{\mu_{2}}\left(1+\frac{zp_{2}}{\mu_{2}}\right)^{-c-1}\right]\sum_{k=1}^{m}\alpha^{m - k}(1-\alpha)^{k-1}\frac{m!}{(m - k)!k!}f_{k}(z , z_{sp})=0\label{lamda2}\end{aligned}\ ] ] where @xmath160 is the lagrange multiplier .
above , @xmath166 denotes the joint distribution of @xmath141 of the channel selected for transmission when all channels are detected busy .
hence , in this case , the transmission channel is chosen among @xmath0 channels .
similarly , @xmath167 denotes the joint distribution when @xmath29 channels are detected idle , and the transmission channel is selected out of these @xmath29 channels . defining @xmath168 and @xmath159 , and solving ( [ lamda1 ] ) and ( [ lamda2 ] )
, we obtain the optimal power policies given in ( [ optimalpolicy1 ] ) and ( [ optimalpolicy2 ] ) .
@xmath152 now , using the optimal transmission policies given in ( [ optimalpolicy1 ] ) and ( [ optimalpolicy2 ] ) , we can express the effective capacity as follows : @xmath169 above , the subscripts @xmath170 and @xmath171 in the expectations denote that the lower limits of the integrals are equal these values and not to zero .
for instance , @xmath172 . until now
, we have not specified the criterion with which the transmission channel is selected from a set of available channels . in ( [ eq : optimaleffectivecap ] ) , we can easily observe that the effective capacity depends only on the channel power ratio @xmath173 , and is increasing with increasing @xmath173 due to the fact that the terms @xmath174 and @xmath175 are monotonically decreasing functions of @xmath173 .
therefore , the criterion for choosing the transmission band among multiple busy bands unless there is no idle band detected , or among multiple idle bands if there are idle bands detected should be based on this ratio of the channel gains . clearly , the strategy that maximizes the effective capacity is to choose the channel ( or equivalently the frequency band ) with the highest ratio of @xmath173 .
this is also intuitively appealing as we want to maximize @xmath95 to improve the secondary transmission and at the same time minimize @xmath94 to diminish the interference caused to the primary users . maximizing @xmath173
provides us the right balance in the channel selection .
we define @xmath176 where @xmath177 is the ratio of the gains in the @xmath178 channel . assuming that these ratios are independent and identically distributed in different channels
, we can express the pdf of @xmath179 as @xmath180^{m-1},\ ] ] where @xmath181 and @xmath182 are the pdf and cumulative distribution function ( cdf ) , respectively , of @xmath173 , the gain ratio in one channel .
now , the expectation @xmath183 , which arises under the assumption that all channels are detected busy and the transmission channel is selected among these @xmath0 channels , can be evaluated with respect to the distribution in ( [ x1 ] ) .
similarly , we define @xmath184 for @xmath185 .
the pdf of @xmath186 can be expressed as follows : @xmath187^{k-1}\quad k=1,2,\dots , m.\ ] ] the expectation @xmath188 can be evaluated using the distribution in ( [ x2 ] ) . finally , after some calculations
, we can write the effective capacity in integral form as @xmath189^{m-1}\left[\frac{\beta_{1}\lambda}{x}\right]^\frac{c}{c+1}dx\nonumber\\&(1-\rho)(1-p_{f})m\int_{\beta_{2}\lambda}^{\infty}f_{\frac{z}{z_{sp}}}(x)\left[\alpha+(1-\alpha)f_{\frac{z}{z_{sp}}}(x)\right]^{m-1}\left[\frac{\beta_{2}\lambda}{x}\right]^{\frac{c}{c+1}}dx+p_{m+2}\bigg\}.\end{aligned}\ ] ]
in this section , we present numerical results for the effective capacity as a function of the channel sensing reliability ( i.e. , detection and false alarm probabilities ) and the average interference constraints . throughout the numerical results
, we assume that qos parameter is @xmath190 , block duration is @xmath191s , channel sensing duration is @xmath192s , and the prior probability of each channel being busy is @xmath193 . before the numerical analysis
, we first provide expressions for the probabilities of operating in each one of the four scenarios described in section [ subsec : state ] .
these probabilities are also important metrics in analyzing the performance .
we have @xmath194 in figure [ fig : fig4 ] , we plot these probabilities as a function of the detection probability @xmath195 for two cases in which the number of channels is @xmath196 and @xmath197 , respectively .
as expected , we observe that @xmath198 and @xmath199 decrease with increasing @xmath0 .
we also see that @xmath200 and @xmath201 are assuming small values when @xmath195 is very close to 1 . note from fig . [
fig : fig1 ] that as @xmath195 approaches 1 , the false alarm probability @xmath202 increases as well .
the analysis in the preceding sections apply for arbitrary joint distributions of @xmath95 and @xmath94 under the mild assumption that the they have finite means ( i.e. , fading has finite average power ) . in this subsection , we consider a rayleigh fading scenario in which the power gains @xmath95 and @xmath94 are exponentially distributed .
we assume that @xmath95 and @xmath94 are mutually independent and each has unit - mean .
then , the pdf and cdf of @xmath173 can be expressed as follows : @xmath203 in fig .
[ fig : fig5 ] , we plot the effective capacity vs. probability of detection , @xmath41 , for different number of channels when the average interference power constraint normalized by the noise power is @xmath204db , where @xmath205 is the noise variance at the primary user .
we observe that with increasing @xmath41 , the effective capacity is increasing due to the fact more reliable detection of the activity primary users leads to fewer miss - detections and hence the probability of scenario 3 or equivalently the probability of being in state @xmath87 , in which the transmission rate is effectively zero , diminishes .
we also interestingly see that the highest effective capacity is attained when @xmath196 .
hence , secondary users seem to not benefit from the availability of multiple channels .
this is especially pronounced for high values of @xmath195 .
although several factors and parameters are in play in determining the value of the effective capacity , one explanation for this observation is that the probabilities of scenarios 1 and 2 , in which the secondary users transmit with power @xmath88 , decrease with increasing @xmath0 , while the probabilities of scenarios 3 and 4 increase as seen in ( [ scenarios ] ) .
note that in scenario 3 , no reliable communication is possible and transmission rate is effectively zero . in fig .
[ fig : fig6 ] , we display similar results when @xmath206db . hence ,
secondary users operate under more stringent interference constraints . in this case
, we note that @xmath207 gives the highest throughput while the performance with @xmath196 is strictly suboptimal . in fig .
[ fig : fig7 ] , we show the effective capacities as a function @xmath208 ( db ) for different values of @xmath0 when @xmath209 and @xmath210 . confirming our previous observation
, we notice that as the interference constraint gets more strict and hence @xmath208 becomes smaller , a higher value of @xmath0 is needed to maximize the effective capacity .
for instance , @xmath197 channels are needed when @xmath211db . on the other hand , for approximately @xmath212db ,
having @xmath196 gives the highest throughput .
above , we have remarked that increasing the number of available channels from which the transmission channel is selected provides no benefit or can even degrade the performance of secondary users under certain conditions .
on the other hand , it is important to note that increasing @xmath0 always brings a benefit to the primary users in the form of decreased probability of interference . in order to quantify this type of gain
, we consider below the probability that the channel selected for transmission is actually busy and hence the primary user in this channel experiences interference : @xmath213 note that @xmath214 depends on @xmath195 and also @xmath202 through @xmath215 .
it can be easily seen that this interference probability @xmath214 decreases with increasing @xmath0 when @xmath216 .
as @xmath0 goes to infinity , we have @xmath217 indeed , in this asymptotic regime , @xmath214 becomes zero with perfect detection ( i.e. , with @xmath218 ) .
note that secondary users transmit ( if @xmath219 ) even when all channels are detected as busy . as @xmath220 , the probability of such an event vanishes .
also , having @xmath218 enables the secondary users to avoid scenario 3 .
hence , interference is not caused to the primary users . in fig .
[ fig : fig3 ] , we plot @xmath214 vs. the detection probability for different values of @xmath0 .
we also display how the false alarm probability evolves as @xmath195 varies from 0 to 1 .
it can be easily seen that while @xmath221 when @xmath196 , a smaller @xmath214 is achieved for higher values of @xmath0 unless @xmath218 . on the other hand ,
as also discussed above , we immediately note that @xmath214 monotonically decreases to 0 as @xmath195 increases to 1 when @xmath0 is unbounded ( i.e. , @xmath220 ) .
nakagami fading occurs when multipath scattering with relatively large delay - time spreads occurs .
therefore , nakagami distribution matches some empirical data better than many other distributions do . with this motivation
, we also consider nakagami fading in our numerical results .
the pdf of the nakagami-@xmath222 random variable @xmath223 is given by @xmath224 where @xmath222 is the number of degrees of freedom .
if both @xmath94 and @xmath95 have the same number of degrees of freedom , we can express the pdf of @xmath225 as follows : @xmath226 note also that rayleigh fading is a special case of nakagami fading when @xmath227 . in our experiments
, we consider the case in which @xmath228 . now
, we can express the cdf of @xmath179 for @xmath228 as @xmath229 in fig .
[ fig : fig8 ] , we plot effective capacity vs. @xmath208 ( db ) for different values of @xmath0 when @xmath209 and @xmath210 . here
, we again observe results similar to those in fig .
[ fig : fig7 ] .
we obtain higher throughput by sensing more than one channel in the presence of strict interference constraints on cognitive radios .
in this paper , we have studied the performance of cognitive transmission under qos constraints and interference limitations .
we have considered a scenario in which secondary users sense multiple channels and then select a single channel for transmission with rate and power that depend on both sensing decisions and fading .
we have constructed a state transition model for this cognitive operation .
we have meticulously identified possible scenarios and states in which the secondary users operate .
these states depend on sensing decisions , true nature of the channels being busy or idle , and transmission rates being smaller or greater than the instantaneous channel capacity values .
we have formulated and imposed an average interference constraint on the secondary users . under such interference constraints and also statistical qos limitations in the form of buffer constraints ,
we have obtained the maximum throughput through the effective capacity formulation .
therefore , we have effectively analyzed the performance in a practically appealing setting in which both the primary and secondary users are provided with certain service guarantees .
we have determined the optimal power adaptation strategies and the optimal channel selection criterion in the sense of maximizing the effective capacity .
we have had several interesting observations through our numerical results .
we have shown that improving the reliability of channel sensing expectedly increases the throughput .
we have noted that sensing multiple channels is beneficial only under relatively strict interference constraints . at the same time , we have remarked that sensing multiple channels can decrease the chances of a primary user being interfered .
v. asghari and s. aissa , rate and power adaptation for increasing spectrum efficiency in cognitive radio networks , " _ ieee international conference on communications , dresden , germany , jun .
14 - 18 , 2009 . _ s. akin and m.c .
gursoy , effective capacity analysis of cognitive radio channels for quality of service provisioning , " _ ieee global communication conference , honolulu , hawaii , nov .
30 - dec . 4 , 2009 . | in this paper , the performance of cognitive transmission under quality of service ( qos ) constraints and interference limitations is studied .
cognitive secondary users are assumed to initially perform sensing over multiple frequency bands ( or equivalently channels ) to detect the activities of primary users .
subsequently , they perform transmission in a single channel at variable power and rates depending on the channel sensing decisions and the fading environment . a state transition model is constructed to model this cognitive operation .
statistical limitations on the buffer lengths are imposed to take into account the qos constraints of the cognitive secondary users . under such qos constraints and limitations on the interference caused to the primary users ,
the maximum throughput is identified by finding the effective capacity of the cognitive radio channel .
optimal power allocation strategies are obtained and the optimal channel selection criterion is identified . the intricate interplay between effective capacity , interference and qos constraints , channel sensing parameters and reliability , fading , and the number of available frequency bands
is investigated through numerical results .
_ keywords : _ channel sensing , cognitive transmission , effective capacity , energy detection , interference constraints , nakagami fading , power adaptation , quality of service constraints , rayleigh fading , state - transition model . 1.65 |
recent years have witnessed significant improvements in the quality and quantity of data from clusters of galaxies in all wavebands , from the x - ray , through optical , to radio wavelengths . together with advances in the theoretical modeling of the physical processes in clusters , these improvements call for new and more accurate methods of data analysis and comparison with various theoretical predictions . in particular , uncovering the 3d structure of clusters is of great interest as it plays an important role in the determination of the hubble constant from x - ray and sz measurements ( see appendix ) , cluster mass and baryon fraction determination , and the underlying galaxy orbit structure . in order to reconstruct the 3d geometry of clusters
researchers normally apply the spherical symmetry assumption , either within a parametric ( cavaliere & fusco - femiano 1976 ) or non - parametric ( fabian 1981 ; and more recently yushikawa & suto 1999 ) framework .
although this simplifying assumption is very useful for many problems , it can lead to significant biases in others , the hubble constant determination ( fabricant et al 1984 ) .
recently , a few studies have developed methods to take advantage of the full two dimensional information retained in the cluster images and extend the modeling of the underlying three dimensional structure to have a more general , aspherical distribution ( zaroubi 1998 ; reblinsky & bartelmann 1999 ) . in particular , zaroubi ( 1998 ; hereafter referred to as paper i ) presented a model - independent method of image deprojection for probing the 3d structure in clusters of galaxies , assuming only that the underlying 3d structure has axial symmetry .
the original testing of this method demonstrated that given images of the cluster in the x - ray , radio and a map of the projected total mass distribution , one could uniquely determine the inclination angle of the system symmetry axis , and determine the underlying 3d structure in the gas and dark matter ( see also grego 2000 ) .
they studied the stability of the inversion method in the context of a specific analytical gas density profile : an axially symmetric , elliptical isothermal model ; which is an extension of the widely used isothermal sphere model .
a few questions arise from the study in paper i : 1 ) is the fundamental assumption of axial symmetry valid for real galaxy clusters ?
2 ) does substructure and dynamical evolution in galaxy clusters ( as predicted to be abundant by hierarchical structure formation models ) permit a robust , unique and stable numerical application of the method ?
3 ) can the cluster symmetry axis inclination angle be uniquely determined for more complex cluster gas / mass density distributions ?
4 ) with realistic expected resolution and signal - to - noise from current observations , can the method be applied ? to probe these questions , we have performed a study of the inversion algorithm on a set of numerically simulated galaxy clusters . the goal is to test the axisymmetric deprojection algorithm , using as input a representative realization of the cluster population .
we would like to emphasize at this stage that testing the applicability of the technique to current data sets is beyond the scope of the current paper .
the method should be applied to each data set while taking into account its unique specifications , resolution , scanning strategy , data acquisition domain ( uv plane _ vs. _ real space ) , etc .
we intend indeed to tackle these issues for specific data sets in a forthcoming paper .
the outline of this study is as follows . in
2 the deprojection method is briefly reviewed , and we develop a new feature of the method that allows an inclination angle - independent determination of the baryon fraction in clusters . in 3 , we briefly describe the input cluster simulations . in 4 , we show in detail the results for a single , prototype simulated cluster . in 5 ,
we apply the method to the full cluster simulation output .
the paper is concluded with a discussion ( 6 ) .
the method for deprojection is described elsewhere ( @xcite ) .
briefly , we deproject cluster images as follows .
we adopt the convention that bold - face symbols denote 3d quantities ( e.g. , @xmath0 ) .
let the observer s coordinate system be defined with the cartesian axes @xmath1 , with the @xmath2 axis aligned with the line of sight .
denote the ( cluster ) source function coordinate system by the axes @xmath3 where the @xmath4-axis is the cluster symmetry axis , forming an angle @xmath5 with respect to the line of sight ( see diagram shown in figure [ fig : config_digram ] ) .
let @xmath6 denote a projected quantity ( image ) of the source function @xmath7 .
the 3d fourier transform ( ft ) of the source function is related to the image fourier transform by @xmath8 } i(x^\prime , y^\prime ) \ , { \rm d } x^\prime \ , { \rm d } y^\prime \nonumber \\ & = & i_{k^\prime}(k_x^\prime , k_y^\prime ) .
\label{eqn : simagereln } \end{aligned}\ ] ] the expression relating the ft of the source function in the observer and cluster rest frames is obtained simple by coordinate rotation , so that @xmath9 where the last equality again holds due to axial symmetry . inverse transforming , we find the desired expression for the source function in real space @xmath10 the implementation of this method is straightforward , assuming for the moment that the inclination angle is known .
the image ft is evaluated at wavevectors @xmath11 and the inverse transform applied . for wavevectors @xmath12 , the argument of
the image ft becomes imaginary defining a cone in k - space known as the `` cone of ignorance '' ( hereafter coi ) .
an interesting property of equation ( [ eqn : maineqn ] ) is that the k - space information in the @xmath13 plane is fully known and independent of the cluster inclination angle .
it is , therefore , straightforward to show , by integrating over eq .
2 with respect to @xmath4 , that the projection over the @xmath14-direction is related to the symmetry axis integrated density , @xmath15 where @xmath16 is ft of @xmath17 . we can also define the z - projected cluster ` mass ' , @xmath18 .
this relation is especially useful in the case of isothermal clusters , where the sz map is proportional to the projected electron number density , hence to the projected gas density , and the surface density map is proportional to the projected total mass density .
therefore equation ( [ eqn : rhoz ] ) provides an angle independent estimate of @xmath19 , and enables evaluation of the ratio @xmath20 and @xmath21 as a function of the distance from the symmetry axis .
note also that since the gas and total mass distributions for an isothermal cluster are roughly proportional to each other in most of the volume ( @xcite ) , except the inner most region , this comparison is almost hubble constant independent .
to test the validity of the deprojection method , we employed a set of outputs from a gas dynamic simulation of the formation of an x - ray cluster .
the simulation is evrard s contribution to the santa barbara cluster comparison project ( @xcite ) .
details of the simulation are provided elsewhere ( @xcite ) , but we very briefly outline the data set here : the underlying structure formation model was cdm , with the g3 transfer function fit of bardeen ( 1986 ) at @xmath22 . the simulation was performed in a flat universe , the initial fluctuation spectrum was given a shape parameter of @xmath23 and normalized such that the present day rms mass fluctuations in a spherical top - hat of radius @xmath24 is @xmath25 .
the cosmological parameters were assigned with as @xmath26 , @xmath27 , and a contribution from the baryons of @xmath28 .
the initial conditions for the simulation were generated with a constrained gaussian random field ( @xcite ) , and galaxy clusters identified with @xmath29 peaks of the field smoothed with a gaussian filter of scale @xmath30 , and centered in a cubic region of side @xmath31 .
the evolution of the gas was followed using the smoothed particle hydrodynamics ( sph ) technique ( @xcite ) to @xmath32 .
final maps of the projected total mass distribution , compton y - parameter , and simplified versions of the x - ray luminosity and x - ray emission weighted temperature were produced .
the projected quantities were calculated from a @xmath33 cube , and computed on a @xmath34 grid , corresponding to a physical scale of @xmath35 at @xmath32 . in the rayleigh - jeans regime ,
the compton y - parameter is related to the sz decrement as @xmath36 .
we have applied this conversion to the simulation output . for the purposes of this test
, we employed three orthogonal projections of a single cluster at four redshifts : @xmath37 .
this , in effect , yields a sample of 12 semi - independent clusters to study with a wide range of dynamical states and physical distributions of the gas and dark matter .
for example , at @xmath38 , the cluster undergoes a major merger event , and thus the outputs at that redshift permit a study of the method while the cluster is in a highly unrelaxed state .
for the first part of this analysis , we concentrate on a single ` prototype cluster ' ; the simulation output at @xmath39 . in figure
[ fig : allproj ] , we display the logarithmically scaled mass surface density , x - ray surface brightness , sz decrement and emission weighted temperature distributions for the three orthogonal projections of the prototype cluster . the dimensions for the surface density , temperature and sz decrement are physical ( @xmath40 and @xmath41 for the first two , while the sz decrement is dimensionless ) .
the x - ray surface brightness was calculated as the line of sight integral @xmath42 and has dimensions @xmath43 . in this section , unless stated otherwise , the images used for deprojection are the y - projection maps of the simulated protocluster . for the prototype cluster ,
the projected gas and total mass distributions are elliptical , with axis ratios @xmath44 . in all of the projections ,
an infalling sublump is evident , approximately @xmath45 ( projected distance ) from the cluster center .
we note that the emission weighted temperature is nearly constant over the innermost @xmath46 , with a value of @xmath47 . in what follows for this specific cluster ,
we treat the gas distribution as being isothermal . using equation ( [ eqn : rhoz ] ) , we calculated the gas to total mass and density ratios for prototype cluster .
the results are shown in figure [ fig : densityratio ] for the y - projection maps . as discussed below ( see [ sec : determining_projected_symmetry_axis ] ) , we do not know apriori if the cluster is prolate or oblate , and so we have calculated the ratios under both assumptions .
however , this is the only unknown ; no knowledge of the inclination angle is required to determine these ratios .
the value of density and mass ratios in the case of this prototype cluster is @xmath48 . under the assumption
the cluster is prolate , this ratio is approximately constant at all radii , whereas in the oblate case , it rises by @xmath49% towards the center .
however , in both cases , at the outermost radii probed by our calculation , the density ratio becomes @xmath50 which is extremely close to the input ratio in the sph / n - body simulation .
the density and mass ratios across the cluster , especially in relaxed clusters , are expected to have an almost constant value , with small radial dependence ( @xcite ) .
therefore , the greater variability shown in the density and mass ratios under the assumption of oblate cluster might be used as an indicator to that the protocluster at hand is most probably prolate .
inspection of the true 3d distribution shows that this is indeed the case .
the same argument can be repeated for all the clusters we have analyzed in this paper ( see figs . [
fig : z00_densityratio ] , [ fig : z03_densityratio ] , [ fig : z06_densityratio ] , and [ fig : z09_densityratio ] ) .
furthermore , the true full 3d distribution clearly shows that the gas radial density profile follows very faithfully the radial mass density profile down to the inner most @xmath51 , interior to this radius the ratio drops down due to the effect of the pressure ( see figs .
10 and 12 in @xcite ) , which explains the drop in the gas to mass ratio at the inner most radii in the prolate case shown in the left panel of figure [ fig : densityratio ] . to demonstrate the utility of the method for the determining the underlying 3d structure
, we proceed in two ways : first , we demonstrate that a robust and stable determination of the inclination angle can be achieved using the sort of information that would be available from observations ( i.e. , x - ray , sz and lensing maps ) . secondly , using our knowledge of the true underlying 3d structure , available
since we are employing simulated clusters , we determine the accuracy with which cumulated quantities ( such as gas or total mass ) can be determined from this method .
we encountered several subtle features of the deprojection procedure which we describe in detail below .
however , the overall procedure is , conceptually at least , straightforward : we first determine the image axis of symmetry , which is the projection of the axis of symmetry in the 3d distribution .
we allow both possibilities that the cluster is either oblate or prolate , and deproject using equation ( [ eqn : maineqn ] ) .
as the inclination angle is unknown , we deproject each image over a wide range of assumed inclination angles , searching for the angle that consistently yields the best reconstructed ratios between the sunyaev - zeldovich , x - ray and the total mass deprojections , as discussed below . to perform the image deprojection
, the first step is to define the ( projected ) symmetry axis .
this turns out to have a couple of subtle complications : first , the simulation data often seems to have , at least qualitatively , a varying axis of symmetry as a function of radius from the image centroid .
secondly , the images become progressively rounder from the surface density , to sz , to x - ray maps respectively how do the axes of symmetry found from each independently compare ? to determine the projected symmetry axis , we proceed as follows . denote the image by @xmath52 $ ] , where @xmath53
$ ] , and @xmath54 is the image size ( @xmath55 for this simulation data ) .
suppose the axis of symmetry is parallel to the @xmath56-axis , passing through the point @xmath57 .
then , for a perfectly axially symmetric object @xmath58 = f[x_c + x , y ] \;\;\ ; \forall \ , x , y .
\label{eqn : axiallysymmetry}\ ] ] given a real image ( which may only be approximately axially symmetric ) , the axis of symmetry is the one that , in some sense , ` best ' satisfies equation ( [ eqn : axiallysymmetry ] ) . to determine this axis empirically
, we scan through a set of points around @xmath59 , ( the starting point is usually chosen to be the coordinates of the object s ` center of mass ' ) and find the line that best satisfies the axially symmetric condition as quantified below . in the case where the pixel counts are uncorrelated , and the model is that the image is axisymmetric with poisson noise
, then we form a @xmath60-like statistic that is minimized when the image is rotated with the axis of symmetry aligned with the y - axis : @xmath61 - f_{\rm rot}[x_{c_n } + x , y ] \right)^2 .
\label{eqn : rsum}\ ] ] one consideration is that a @xmath62 estimator is only strictly valid in the limit that the errors follow a gaussian distribution .
however , in practise , x - ray observations are counting experiments , often with low signal - to - noise , and such estimators are not valid .
an alternative is to consider a robust m - estimator ( @xcite ) .
if we assume that the that the errors are distributed exponentially , then the maximum likelihood estimator for the axis angle is determined by minimizing over the sum @xmath63 - f_{\rm rot}[x_{c_n } + x , y ] \
, \right|^{1/2 } .
\label{eqn : msum}\ ] ] we also considered an alternative , and less formal , approach to determining the symmetry axis .
this was partially motivated by the shortcomings of the @xmath62 estimator for poisson errors .
furthermore , since the x - ray and , to a lesser extent , the radio emission and mass density , are typically quite centrally concentrated , in the poisson model for the noise , there is little contribution to the @xmath64 and @xmath65 sums from data at large radii from the cluster center , even if the difference in counts between symmetric pixels is large .
thus , the above estimators will be dominated by terms only from the central region of the cluster . to enable us to be sensitive to isophote rotation
, we have adopted relative difference between symmetric pixels , which will give equal weight to _ all _ image pixels @xmath66 - f_{\rm rot}[x_{c_n } + x , y]}{f_{\rm rot}[x_{c_n } - x , y ] + f_{\rm rot}[x_{c_n } + x , y ] } \right|.\ ] ] a ` good fit ' for the symmetry axis will have @xmath67 , while poor fits will have @xmath68 .
we originally were interested to see if we could find an optimal estimator of the symmetry axis , based on equation ( [ eqn : axiallysymmetry ] ) .
however , the brute force implementation of the sums to compute @xmath64 , @xmath65 and @xmath69 are computationally very fast , and hence we have not pursued this further . in calculating all of these statistics to find the image axis of symmetry
, we have allowed the point @xmath59 to be a free parameter in the fit .
this becomes a highly non - linear minimization problem , and we have used the simplex method ( @xcite ) to simultaneously determine the axis angle and centroid . in what follows , we have employed the @xmath69 to determine the projected image axis .
this choice was made somewhat arbitrarily ; our rationale was that since we are primarily interested in determining the long wavelength cluster features , we prefer to give increased weight to the image counts at large radii from the cluster center . in practice , this choice seems to work very well for recovering the cluster 3d structure , as we demonstrate in what follows below .
however , we do note that using either the @xmath64 or @xmath65 statistics also give good results , and we do not have a definitive recommendation as to which to use in analyses of real data ( in fact , during our preliminary work with the simulated cluster sample for this study , we used all three estimators and a similar procedure would be prudent when applying the method to real data ) .
one final complication arises in that apriori we do not know if the cluster is prolate or oblate .
that is , in general , there will be two axes that best satisfy the axial symmetry condition corresponding to the case that the cluster is oblate or prolate . in the context of the deprojection and
what follows , we use both axes and test via a comparison of the resulting 3d structure under each hypothesis which is valid . having defined the image axis of symmetry ,
the next operation in the deprojection algorithm requires filling the coi of each deprojection . in paper
i , a simple extrapolation scheme was used to fill the coi .
here we have explored somewhat more sophisticated extrapolation methods , such as fitting an elliptical isothermal model to the actual ft of the image outside the coi , and using the model fit to extrapolate into the coi .
we find that our results are fairly insensitive to the precise algorithm applied to fill the coi , primarily because the information lost inside the coi is restricted to moderately high spatial frequencies .
whatever scheme we apply to fill the coi affects most the small scale details of the 3d densities we infer , while the large scale features are less affected .
this is clearly seen in figure [ fig : coi ] where we plot the k - space structure of the sz , x - ray and surface density maps , where we have assumed for the moment that the inclination angle is @xmath70 ( this simply defines the region occupied by the coi .
we motivate this choice of angle in [ sec : determining_theta ] ) .
most of the long wavelength power is outside the coi , which tends to dominate k - space primarily at fairly high spatial frequencies .
still , in order to reduce the effect of ringing by the discontinuities at the edge of the coi , we need to do some sort of extrapolation into this regime .
we display two alternatives here : first we simply perform a linear extrapolation into the coi with the amplitude fixed by the value at the cone boundary .
this is shown by the solid line in figure [ fig : coi ] .
secondly , we note that for this cluster , the image ft of the sz and x - ray images are reasonably well fit by the elliptical isothermal gas density model , and we have used this model to extrapolate the observed image ft into the coi with the dashed line in figure [ fig : coi ] .
this prescription seems quite reasonable for the sunyaev - zeldovich and x - ray surface brightness maps , while the fit for the surface density map is not as successful .
this is somewhat expected as the former ( which probe the gas density distribution ) are smoother and not as sensitive to the details manifest in the cluster s substructure as the surface density map .
having a prescription to fill the coi , we are now in position to do the deprojection , compare the results from the various projections , and determine the cluster inclination angle .
we first consider the comparison of the sz _ vs. _ x - ray deprojections .
we deproject each image independently ( with the same inclination angle ) , solve for the 3d gas density , and compare the results from each deprojection . in paper
i , we demonstrated that in order to determine the inclination angle from a comparison of the sunyaev - zeldovich and x - ray maps , the _ shape _ alone of the density profiles from each image is insufficient as the method , for any given angle , produces the same profile shape from both maps . the only way to determine the angle is therefore from the ratio of the _
amplitudes_. this ratio is used to determine the inclination angle ( where we use our knowledge , for the moment , that @xmath71 ) yielding a best fit value of @xmath72 and @xmath73 for prolate and oblate cluster respectively .
figure [ fig : szvsxray ] shows the gas density distribution as inferred from the deprojection of the sz and x - ray maps for prolate and oblate clusters using the appropriate best fit angle for each case .
the agreement between the two , in terms of both shape and amplitude , is very good ( within @xmath74 difference at the center ) .
the comparison between the sunyaev - zeldovich and surface density maps is somewhat more complicated , as we need to apply the hydrostatic equation to relate the dark matter and gas distributions @xmath75 where @xmath76 , @xmath77 and @xmath78 are the gas and total mass densities and the cluster s temperature respectively .
the gas is assumed to be an ideal gas .
for an isothermal cluster this equation becomes very simple , @xmath79 , where @xmath80 is a ( known ) numerical constant .
note that in the limit of @xmath81 equation ( [ eqn : hydro ] ) is independent of the amplitude of the gas density , @xmath76 and therefore is sensitive to the shape of the gas profile .
this forms an attractive supplement to the sz versus x - ray comparison , which depends only on the amplitude . operationally , to compare the sz and surface density deprojections we proceed in a similar fashion as with the sz and x - ray maps : the sz and surface density maps
are deprojected for many angles assuming oblate and prolate structures . for each angle
the gas density is inserted in equation ( [ eqn : hydro ] ) and the result is compared with the total density profile by a @xmath62-like statistic .
the best fit between the two again occurs at the angles @xmath70 and @xmath82 for the prolate and oblate assumptions respectively .
these comparisons are shown in figure [ fig : dmvsgas ] .
since we have three projections of the same underlying cluster , we can make a more sophisticated comparison .
we deproject the y - projections and _ predict _ the two other orthogonal projections , searching for the inclination angle that provides the best fit when compared with the true x- and z - axis projections for the sz , x - ray and surface density images .
we again find an inclination angle of @xmath72 for the prolate case and @xmath82 for the oblate case .
a comparison between the projected profiles and the real projections are shown in the left panels of figures [ fig : projection_lens ] , [ fig : projection_sz ] and [ fig : projection_xray ] .
it is evident from those figures that the reconstruction works very well apart from the innermost region where the amplitude at the very center ( @xmath83 ) is underestimated by @xmath84 .
furthermore , due to heavy reliance on the elliptical isothermal gas density distribution model for filling the coi , the recovered 3d cluster distribution is somewhat over- elongated . for an inclination angle of @xmath85
the coi covers most of k - space and one is forced to heavily rely on the assumed model to compensate for the lost information . as another example we deproject the z - projection image which has , for a prolate cluster , a rather large inclination angle , @xmath86 , and therefore has a negligible coi .
the reconstructions from the z - projection maps are shown in the left panels of figures [ fig : projection_lens ] , [ fig : projection_sz ] and [ fig : projection_xray ] . due to the minor coi extrapolation in this case the cluster center and geometry
are better recovered and the other two projections are well predicted , even in the innermost region of the cluster .
figure [ fig : obl_projection ] shows the same comparison shown in the left panels of figure [ fig : projection_lens ] but for an oblate cluster . in this case
the reconstructed x - projection is rather poorly predicted , supporting our contention regarding the cluster prolateness .
since we have the full 3d information from the simulations , we can robustly test the accuracy of the deprojection by comparing the deprojection with the true 3d gas or total mass distributions . in figure
[ fig : chi2vstheta ] , we deproject the x - ray and sz images as a function of inclination angle . we transform the deprojected x - ray image to calculate the gas density from the deprojection , and compare with the true gas density by forming the normalized absolute difference : @xmath87 where the @xmath88 is the true ( axi - symmetrized ) density field in the simulation . in the left and middle panels of figure [ fig : chi2vstheta ]
, the dashed line displays the normalized absolute difference between the inferred and true gas densities as a function of inclination angle . for comparison
, we also display ( solid line ) the difference between the gas density inferred from the x - ray deprojection versus that from the sz this mimics the procedure one would do in practice with real data , as discussed in
[ sec : determining_theta ] . for this cluster ,
the agreement between the two is superb , especially under the assumption of a prolate cluster , demonstrating the reliability of the inclination angle determination ( within @xmath89 ) even in the case of a very wide coi .
a useful way to quantify the quality of the deprojection is to is see how well the radial profiles as calculated from the full and deprojected 3d distributions compare . in figure [ fig : spherical ] we calculate the radial profiles of the gas distribution and dark matter distribution . in both cases
the reconstruction underestimated the density at the inner most center of the cluster but it yields a very accurate profile at radii @xmath90 for the gas density profile and @xmath91 for the dark matter density profile .
the discrepancy between the real and reconstruction radial profiles in the innermost region , especially in the dark matter radial profile , is primarily due to the loss of the high frequency information in the coi , coupled with the finite resolution of the ` observed ' maps .
however , the _ total _ gas and mass differences are only of the order of few percent . in the full deprojection results shown until now
, we have filled the coi by extrapolating an isothermal ellipsoidal profile that best fits the data outside the coi to within the coi .
here , we explore the effect of two different schemes for filling the coi . in the first scheme we simply let the values within the coi to smoothly ( exponentially )
drop to zero .
the second scheme is to simply perform a linear extrapolation into the coi with the amplitude fixed by the value at the cone boundary ( see figure [ fig : coi ] ) .
the general effect of first scheme is twofold .
first , it results in a decrease in the amplitude of reconstructed source function , with the magnitude of the drop depending on the size of the coi .
second , it produces a double lobed shape distribution .
as well , the second scheme results in a decrease in the amplitude of the reconstructed source function , and in more boxy shaped structures .
figure [ fig : coi_test ] shows the _ predicted _ images obtained by projecting the 3d reconstructed sz source function into x- , y- , and z - directions .
here we use the y - projected sz as the input image for the algorithm .
the inclination angle here is @xmath70 which means that the coi is quite wide .
as expected , the re - projected y - image has the same quality as the comparable image ( first column and second row ) in figure [ fig : projection_sz ] .
the other two re - projected images show in both cases a drop in the amplitude and in the case of `` constant filling '' the reconstruction has a more boxy features while the `` zero filling '' results in a double lobed reconstruction .
we have demonstrated thus far that our procedure for filling the coi and comparing the x - ray and sz maps is sufficient to uniquely determine the inclination angle and hence determine the underlying 3d density structure . however , our calculations thus far have employed essentially infinite spatial resolution , and have been noise - free . in order to test
if the method will be viable with realistic noise and instrumental response , we performed the following simple simulation : the sz maps have been degraded to obtain @xmath92 spatial resolution and peak @xmath93 similar to the resolution and sensitivity of the sz observation attained at the ovro and bima arrays ( @xcite ) . similarly , the x - ray images have been degraded to spatial resolution of @xmath94 , and peak @xmath95 , mimicking the resolution and sensitivity of chandra for a rich cluster of galaxies .
the right panel of figure [ fig : chi2vstheta ] shows the angle determined from the x - ray _ vs. _ sz maps comparison .
the maps are initially denoised with a wavelet denoising algorithm ( see paper i ; brosch & hoffman 1999 ; hoffman 2000 ) and then the deprojected x - ray and sz images are compared as a function of inclination angle .
the best fit inclination angle in this case is @xmath96 which is within @xmath97 from the angle obtained from noiseless images ; this is an excellent agreement that shows the potential of this method when applied to real data .
in figure [ fig : allproj_z03 ] we display the logarithmically scaled mass surface density , x - ray surface brightness , sz decrement and emission weighted temperature distributions for the three orthogonal projections of the cluster at @xmath98 . as before , the dimensions for the surface density , temperature and sz decrement are physical ( @xmath40 and @xmath41 for the first two , while the sz decrement is dimensionless ) , while we recall that the x - ray surface brightness was calculated as the line of sight integral @xmath99 and has dimensions @xmath100 . figures [ fig : allproj_z06 ] and [ fig : allproj_z09 ] show the same plots for the cluster at @xmath38 and @xmath101 respectively . the merger history of the cluster can be traced through this series of simulation outputs . at @xmath102 ,
the central mass concentration is in place , while two large sublumps are infalling at a projected co - moving distance of @xmath103 from the cluster center . at @xmath38 ,
the entire system undergoes a major merger event .
the three main mass concentrations are most clearly seen in the cores of the x- and y - projections ( figures [ fig : allproj_z09 ] and [ fig : allproj_z06 ] ) , while the morphology of both the gas and dark matter in the z - projection becomes highly elliptical ( figure [ fig : allproj_z03 ] ) .
this particular simulation output provides the most stringent test for the deprojection method : the system is clearly far away from dynamical equilibrium and certainly is very poorly described by a spherical model for the underlying gas and dark matter distributions .
nevertheless , as we shall demonstrate , the system is fairly well described as axially symmetric ; the symmetry is qualitatively evident in the sz and x - ray images , and , to a lesser extent , in the surface mass distributions .
using equation ( [ eqn : rhoz ] ) , we calculated the gas to total mass and density ratios for the full suite of cluster outputs .
the results are shown in figures [ fig : z00_densityratio ] , [ fig : z03_densityratio ] , [ fig : z06_densityratio ] , and [ fig : z09_densityratio ] for the simulation outputs at @xmath104 respectively . in all cases ,
the method recovers very well the simulation input baryon fraction .
the radial dependence of the gas to total mass ratio is flatter under the assumption of prolateness in the underlying 3d structure .
therefore , one can possibly use the expected small variability in gas to mass ratio to determine whether the cluster is prolate or oblate .
in all the cases we examined , the gas to mass ratio shows less variability when its shape is assumed to be prolate , which is in total agreement with the real distribution of the gas and mass in those clusters .
the inclination angle determination from the cluster sz _ vs. _ x - ray images at the high redshifts is somewhat more complicated .
this complication stems from the fact that at higher redshifts the cluster is not very relaxed , especially at @xmath105 where the cluster undergoes a major merger , rendering the connection between the images and the temperature uncertain .
nevertheless , we have attempted to reconstruct assuming a constant cluster temperature , chosen to be the mean temperature in the region around the cluster center with ` intensities ' larger the @xmath106 of the image maximum density .
we note that the scatter around that mean temperature in the chosen region is about @xmath107 at worst .
figure [ fig : angle_det ] shows the inclination angle as determined from a comparison of the sz and x - ray images as viewed from the x , y and z - projections and at four different redshifts _ vs. _ the inclination angle as determined from the actual 3d distribution .
the agreement between the two is very good .
= 0.5 obviously , filling the coi with an elliptical isothermal density becomes unrealistic for the @xmath105 case , where a major merger is taking place . in principle , one can vary the density model adopted for filling the coi .
here we choose to stay with the isothermal model with the hope of obtaining a reasonable reconstruction the cluster central region .
figures [ fig : projection_sz3 ] , [ fig : projection_sz6 ] and [ fig : projection_sz9 ] shows how well the method performs in the various redshifts when the z - projection images are used for reconstruction .
note the reconstruction is reasonable in all cases with the exception of the @xmath105 case where the details are different .
we have tested a method for deprojecting realistic galaxy cluster x - ray , sz decrement and mass surface density maps , assuming only that the underlying gas / mass distributions are axially symmetric .
conceptually , the deprojection procedure is straightforward : we first determine the image axis of symmetry , employing three different estimators accounting for the poisson nature of the noise , and a technique giving increased weight to counts at large radii from the cluster center .
having a solution for the projected symmetry axis , we deproject the sz , x - ray and surface mass distributions using equation ( [ eqn : maineqn ] ) , allowing for the possibility that the cluster is either oblate or prolate .
as the inclination angle is the primary unknown , we deproject each image over a wide range of assumed inclination angles , searching for the angle that consistently yields the best reconstructed ratios between the sunyaev - zeldovich , x - ray and the total mass deprojections . at each inclination angle
, we fill the information lost in the coi by either an elliptical isothermal model fitted to the data outside the coi , or a linear extrapolation into the coi with the amplitude fixed by the value at the cone boundary .
the results are fairly insensitive to the precise algorithm applied to fill the coi , primarily because the information lost inside the coi is restricted to moderately high spatial frequencies . for this study
, we demonstrated this procedure in depth for a single , prototype cluster , and then applied the method to the entire sample .
we demonstrated the goodness and robustness of the inclination angle determination in several ways : we compared the 3d gas density distribution as inferred from the sz and x - ray deprojection , using the appropriate best fit angle for each case .
the agreement between the two , in terms of both shape and amplitude , is very good ( with a @xmath74 difference at the center ) .
we also determined the inclination angle by calculating the cumulated , normalized absolute difference between the inferred and true gas densities as a function of inclination angle ( see figure [ fig : chi2vstheta ] ) .
the results agreed with our previous determinations to an accuracy of @xmath89 , demonstrating the reliability of the inclination angle determination even in the case of a very wide coi . with the prototype cluster ,
we mimic observations of a rich galaxy cluster by chandra ( x - ray ) and bima or ovro ( sz ) by degrading the resolution of the simulated data and adding noise .
we again compared the deprojections of the two maps to determine the inclination angle and found @xmath96 ( assuming a prolate cluster ) which is within @xmath97 from the angle obtained from noiseless images ; this is an excellent agreement that shows the potential of this method when applied to real data .
we note here , that the purpose of this exercise is to show that the method has the potential of reproducing the 3d structure from the current data sets , this however does not mean that we have tested the application of the method to those data sets .
the application of the method to each of those data sets and its unique specifications should be tested much more extensively , something we intend to carry out in a forthcoming study .
we also note , that since the current method is based on fourier slice theorem it is well suited to analyze interferometric data , like the data obtained by the bima and ovro experiments , where the raw data is given in the fourier space u - v plane . for this kind of data one can for example , manipulate the x - ray or lensing data and transform them to the fourier u - v plane and make the comparison with the sz raw data at the wavenumbers covered in this plane which has the advantage of avoiding the u - v plane extrapolation needed in order to obtain a real space sz image . finally , with the prototype cluster we tested the accuracy of the deprojection by comparing the true and inferred 3d radial profiles of the gas and total mass distributions ( figure [ fig : spherical ] ) .
we found that the reconstruction underestimated the density at the innermost center of the cluster but it yields a very accurate profile at radii @xmath90 for the gas density profile and @xmath91 for the dark matter density profile .
the discrepancy between the real and reconstruction radial profiles in the innermost region , especially in the dark matter radial profile , is primarily due to the loss of the high frequency information in the coi .
the _ total _ , estimated over the whole the simulation volume , gas and mass differences are only of the order of few percent .
we also developed a feature of the method that provides a direct determination of the baryon fraction in clusters , independent of the cluster inclination angle .
the only uncertainty comes from whether the underlying cluster gas density distribution is assumed to be prolate or oblate . for all of the clusters studied here
, we note that the relative variations in the radial dependence for the baryon fraction for the prolate case are much smaller than if the cluster is oblate ( see figures [ fig : z00_densityratio ] , [ fig : z03_densityratio ] , [ fig : z06_densityratio ] , and [ fig : z09_densityratio ] ) .
the shape of the baryon fraction radial profile provides an interesting possibility for discriminating between the prolate / oblate hypotheses : due to the expected strong correlation between the gas and dark matter distributions , one expects their density and mass ratios to show relatively small variability as a function of radius from the cluster center .
indeed , for low- to intermediate redshift clusters @xmath108 , the baryon fraction has a relatively slight radial dependence @xmath109 ( @xcite ) . in conclusion
, we have demonstrated that : 1 ) the fundamental assumption of axial - symmetry is a reasonable assumption when applied to a realistic subset of the general cluster population .
2 ) under the assumption of axial symmetry , the method works very well for the purpose of recovering the 3d structure .
3 ) for a reasonable coi extrapolation scheme , the deprojection is robust , stable and unique even when most of k - space information has been lost in the projection .
4 ) the method is applicable for deprojecting cluster maps throughout the evolutionary history .
bardeen , j. m. , bond , j. r. , kaiser , n. , & szalay , a. s. 1986 , apj , 304 , 15 brosch , n. , & hoffman , y. 1999 , mnras , 305 , 241 carlstrom , j. e. ; joy , m. ; grego , l. , 1995 , apjl , 456 , l75 .
cavaliere , a. , & fusco - femaino , r. 1976 , a&a , 49 , 137 ettori , s. , & fabian , a. c. 1999 , mnras , 305 , 834 evrard , a. e. 1988 , mnras , 235 , 911 fabian , a. c. , hu , e. m. , cowie , l. l. , & grindlay , j. 1981 , apj , 248 , 47 fabricant , d. , rybicki , g. , gorenstein , p. , 1984 ,
apj , 286 , 186 frenk , c. s. , white , s. d. m. , bode , p. , bond , j. r. , bryan , g. l. , cen , r. , couchman , h. m. p. , evrard , a. e. , gnedin , n. , jenkins , a. , khokhlov , a. m. , klypin , a. , navarro , j. f. , norman , m. l. , ostriker , j. p. , owen , j. m. , pearce , f. r. , pen , u .- l . , steinmetz , m. , thomas , p. a. , villumsen , j. v. , wadsley , j. w. , warren , m. s. , xu , g. , & yepes , g. 1999 , apj , 525 , 554 grego , l. , carlstrom , j. e. , joy , m. k. , reese , e. d. , holder , g. p. , patel , s. , cooray , a.r . , & holzapfel , w. l. , 2000 , apj , 539 , 39 hoffman , y. , & ribak , e. 1991 , apj , 380 , l5 mewe , r. , gronenschild , e. , & oord , van den g. 1985 , a&as , 62 , 197 mewe , r. , lemen , j. r. , & oord , van den g. 1986 , a&as , 65 , 511 nelder , j. a. , & mead , r. 1965 , computer journal , 7 , 308 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 1992 , in numerical recipes ( cambridge : cambridge university press ) , 694 yoshikawa , k. , & suto , y. 1999 , apj , 513 , 549 .
reblinsky , k. , & bartelmann , m. 1999 , in the proceedings of `` gravitational lensing : recent progress and future goals '' , boston university , editors t.g .
brainerd and c.s .
zaroubi , s. , squires , g. , hoffman , y. , & silk , j. 1998 , apj , 500 , 87l
the temperature decrement ( the sunyaev - zeldovich effect ) the rayleigh - jeans regime due to upscattering of microwave background photons by the diffuse cluster electron gas in is given by @xmath110 where @xmath111 is the thompson cross section , @xmath112 the boltzmann constant , @xmath113 is the electron mass , and @xmath114 denotes the 3d electron gas distribution , with temperature @xmath115 . these expressions can be rewritten in a general form , factoring out the normalization and 3d structural forms so that gas density an temperature are given by @xmath118 and @xmath119 where @xmath120 is the cluster angular diameter distance and @xmath121 . for galaxy clusters , the dominant emission mechanism is bremsstrhlung in the limit of an optically thin plasma , so that @xmath122 and is a well - known function ( @xcite ; @xcite ) .
thus we can write @xmath123 and the sz decrement can be written @xmath124 and @xmath125 we can slightly generalize the standard @xmath127-model , to allow for a more complicated 3d geometry .
consider as a toy - model a @xmath132 , isothermal cluster with a gas density profile given by @xmath133 without loss of generality , let us suppose the cluster z - axis forms an angle @xmath134 with respect to the line of sight . in this model ,
the sunyaev - zeldovich decrement is given by @xmath135^{-1/2 } \nonumber \\ & = & 2.5 \times 10^{-4 } \ ; c^\prime \ ; \biggl ( \frac { r_c } { 100 \ ; { \rm kpc } } \biggr ) \ ; \biggl ( \frac{n_0 } { 0.01 \ ; { \rm cm}^{-3 } } \biggr ) \ ; \biggl ( \frac{t_x } { 10 \ ; { \rm kev } } \biggr ) \nonumber \\ & \ ; & \;\;\;\;\;\ ; \times \biggl [ 1 + \frac{x^2}{r_c^2a^2 } + \frac{y^2}{r_c^2 ( c^2\sin^2\theta + b^2\cos^2\theta ) } \biggr]^{-1/2 } \nonumber\end{aligned}\ ] ] where @xmath136 it is immediately clear the error one makes when fitting a spherical isothermal @xmath127-model when the true underlying gas density follows this triaxial model . in the special case that @xmath138 , then from equation ( [ eqn : da ] ) , we find @xmath139 , where @xmath140 is the estimated value of the hubble constant ( see figure [ fig : h0fit ] ) .
we determine the error in the general case empirically . the value of @xmath141 extracted from sz and x - ray observations assuming a spherical @xmath127 model will depend on the fitted values of the normalizations , @xmath127 and @xmath142 . | a general method of deprojecting two - dimensional images to reconstruct the three dimensional structure of the projected object specifically x - ray , sunyaev - zeldovich ( sz ) and gravitational lensing maps of rich clusters of galaxies assuming axial symmetry ( @xcite ) , is considered . here
we test the applicability of the method for realistic , numerically simulated galaxy clusters , viewed from three orthogonal projections at four redshift outputs .
we demonstrate that the assumption of axial symmetry is a good approximation for the 3d structure in this ensemble of galaxy clusters . applying the method
, we demonstrate that a unique determination of the cluster inclination angle is possible from comparison between the sz and x - ray images and , independently , between sz and surface density maps .
moreover , the results from these comparisons are found to be consistent with each other and with the full 3d structure inclination angle determination .
the radial dark matter and gas density profiles as calculated from the actual and reconstructed 3d distributions show a very good agreement . the method is also shown to provide a direct determination of the baryon fraction in clusters , independent of the cluster inclination angle .
# 1#2=.45 = .45 # 1#2=.95=.95 # 1#2#3=.3 = .3 = .3 |
counting the number of solutions for constraint satisfaction problem , denoted by # csp , is a very important problem in artificial intelligence ( ai ) . in theory ,
# csp is a # p - complete problem even if the constraints are binary , which has played a key role in complexity theory . in practice ,
effective counters have opened up a range of applications , involving various probabilistic inferences , approximate reasoning , diagnosis , and belief revision . in recent years , many attentions have been focused on counting a specific case of # csp , called # sat . by counting components , bayardo and pehoushek presented an exact counter for sat , called relsat [ 1 ] . by combining component
caching with clause learning together , sang et al .
created an exact counter cachet [ 2 ] .
based on converting the given cnf formula into d - dnnf form , which makes the counting easily , darwiche introduced an exact counter c2d [ 3 ] . by introducing an entirely new approach of coding components
, thurley addressed an exact counter sharpsat [ 4 ] . by using more reasoning , davies and bacchus addressed an exact counter # 2clseq [ 5 ] .
besides the emerging exact # sat solvers , wei and selman presented an approximate counter approxcount for sat by using markov chain monte carlo ( mcmc ) sampling [ 6 ] .
building upon approxcount , gomes et al . used sampling with a modified strategy and proposed an approximate counter samplecount [ 7 ] .
relying on the properties of random xor constraints , an approximate counter mbound was introduced in [ 8 ] . using sampling of the backtrack - free search space of systematic
sat solver , sampleminisat was addressed in [ 9 ] .
building on the framework of samplecount , kroc et al . exploited the belief propagation method and presented an approximate counter bpcount [ 10 ] . by performing multiple runs of
the minisat sat solver , kroc et al .
introduced an approximate counter , called minicount [ 10 ] .
recently , more efforts have been made on the general # csp problems .
for example , angelsmark et al . presented upper bounds of the # csp problems [ 11 ] .
bulatov and dalmau discussed the dichotomy theorem for the counting csp [ 12 ] .
pesant exploited the structure of the csp models and addressed an algorithm for solving # csp [ 13 ] .
dyer et al . considered the trichotomy theorem for the complexity of approximately counting the number of satisfying assignments of a boolean csp [ 14 ] .
yamakami studied the dichotomy theorem of approximate counting for complex - weighted boolean csp [ 15 ] .
though great many studies had been made on the algorithms for the # csp problems , only a few of them related to the # csp solvers .
gomes et al .
proposed a new generic counting technique for csps building upon on the xor constraint [ 16 ] . by adapting backtracking with tree - decomposition , favier et al .
introduced an exact # csp solver , called # btd [ 17 ] .
in addition , by relaxing the original csp problems , they presented an approximate method approx # btd [ 17 ] . in this paper
, we propose a new type of method for solving # csp problems .
the method derives from the partition function based on introducing the free energy and capturing the relationship of probabilities of variables and constraints .
when computing the number of the solutions of a given csp formula according to the partition function , we require the marginal probabilities of each variable and each constraint to plug into the partition function . in order to obtain the marginal probabilities , we employ the belief propagation ( bp ) because it can organize a computation that makes the marginal probabilities computing tractable and eventually returns the marginal probabilities .
in addition , unlike the counter bpcount using the belief propagation method for obtaining the information deduced from solution samples in samplecount , we employ the belief propagation method for acquiring information for partition function .
this leads to two differences between bpcount and our counter .
the first one is the counter bpcount requires to iteratively perform the belief propagation method and repeatedly obtain the marginal probabilities of each variable on the simplified sat formulae ; while our counter carries out the belief propagation method only once , which spends less cost .
the second one is that the two counters obtaining the exact number of solutions depending on different circumstances .
the counter bpcount needs the corresponding factor graphs of the simplified sat formulae all have no cycles ; while our counter only needs the factor graph of the given csp formula has no cycle , which meets easily .
our experiments reveal that our counter for csp , called pfcount , works quite well .
we consider various hard instances , including the random instances and the structural instances .
for the random instances , we consider the instances based on the model rb close to the phase transition point , which has been proved the existence of satisfiability phase transition and identified the phase transition points exactly .
with regard to the random instances , our counter pfcount improves the efficiency tremendously especially for instances with more variables .
moreover , pfcount presents a good estimate to the number of solutions for instances based on model rb , even if the instances scales are relatively large .
therefore , the effectiveness of pfcount is much more evident especially for random instances . for the structural instances , we focus on the counting problem based on graph coloring .
the performance of pfcount for solving structural instances is in general comparing with the random instances because pfcount sometimes ca nt converge .
however , once pfcount can converge , it can estimate the number of the solutions of instances efficiently . as a whole , pfcount is a quite competitive # csp solver .
a constraint satisfaction problem ( csp ) @xmath0 is defined as a pair @xmath1 , where @xmath2 is a set of variables and @xmath3 is a set of constraints defined on _
v_. for each variable @xmath4 in _ v _ , the domain @xmath5 of @xmath4 is a set with @xmath6 values ; the variable @xmath4 can be only assigned a value from @xmath5 . a constraint _
c _ , called a _ k_-ary constraint , consists of _ k _ variables @xmath7 and a relation @xmath8 , where @xmath9 , @xmath10 , ... , @xmath11 are distinct .
the relation @xmath12 specifies all the allowed tuples of values for the variables @xmath7 which are compatible with each other .
the variable configuration of a csp @xmath0 is @xmath13 that assigns each variable a value from its domain .
a solution to a constraint @xmath14 is a variable configuration @xmath15 that sets values to each variable in the constraint @xmath14 such that @xmath16 .
we also say that the variable configuration @xmath17 satisfies the constraint @xmath14 .
a solution to a csp @xmath0 is a variable configuration such that all the constraints in _ c _ are satisfied . given a csp @xmath0 ,
the decision problem is to determine whether the csp @xmath0 has a solution .
the corresponding counting problem ( # csp ) is to determine how many solutions the csp @xmath0 has .
a csp @xmath0 can be expressed as a bipartite graph called factor graph ( see fig .
1 ) . the factor graph has two kinds of nodes , one is variable node ( which we draw as circles ) representing the variables , and the other is function node ( which we draw as squares ) representing the constraints .
a function node is connected to a variable node by an edge if and only if the variable appeares in the constraint . in the rest of this paper
, we will always index variable nodes with letters starting with _
i _ , and factor nodes with letters starting with @xmath14 .
in addition , for every variable node _ i _ , we will use _ v_(_i _ ) to denote the set of function nodes which it connects to , and _
v_(_i_)@xmath18 to denote the set _
v_(_i _ ) without function node @xmath14 . similarly ,
for each function node @xmath14 , we will use _
v_(@xmath14 ) to denote the set of variable nodes which it connects to , and _
i _ to denote the set _
v_(@xmath14 ) without variable node _ i_. .
the csp @xmath0 is encoded as @xmath20 , where _
v_(@xmath9 ) = \{1 , 2 , 3 } , _
v_(@xmath10 ) = \{1 , 4 , 5 , 6 } , _
v_(@xmath21 ) = \{1 , 7 , 8}. ]
in this section , we present a new approximate approach , called pfcount , for counting the number of solutions for constraint satisfaction problem .
the approach derives from the partition function based on introducing the free energy and capturing the relationship of probabilities of variables and constraints . in the following , we will describe the partition function in details . in this subsection
, we present a partition function for counting the number of solutions for csp .
the partition function is an important quantity in statistical physics , which describes the statistical properties of a system .
most of the aggregate thermodynamic variables of the system , such as the total energy , free energy , entropy , and pressure , can be expressed in terms of the partition function . to facilitate the understanding ,
we first describe the notion of the partition function . given a system of _ n _ particles , each of which can be in one of a discrete number of states , i.e. , @xmath22 , and a state of the system _ x _ denoted by @xmath13 , i.e.
i__th particle @xmath4 is in the state @xmath23 , the partition function in statistical physics is defined as @xmath24 where _
t _ is the temperature , _ e_(_x _ ) is the energy of the state _ x _ , and _ p_(_x
_ ) is the probability of the state _
x_. in this paper , we focus on the partition function that the temperature _ t _ is assigned to 1 . since the partition function is also used in probability theory , in the following we will learn the partition function from the probability theory .
given a csp @xmath0 and a variable configuration @xmath13 of @xmath0 , the partition function in probability theory is defined in equation ( 2 ) .
@xmath25 where _
p_(_x _ ) is the the joint probability distribution , function @xmath26 is a boolean function range \{0 , 1 } , which evaluates to 1 if and only if the constraint @xmath14 is satisfied , evaluates to 0 otherwise ; and _ m _ is the number of constraints .
based on equation ( 2 ) , the joint probability distribution _ p_(_x _ ) over the _
n _ variables can be expressed as the follows .
@xmath27 because the construction of the joint probability distribution is uniform over all variable configurations , _ z _ is the number of solutions of the given csp @xmath0 .
therefore , # csp can be solved by computing a partition function . in the following
, we will propose the derivation of the partition function . in order to present a calculation method to compute the partition function
, we introduce the variational free energy defined by @xmath28 where _
e_(_x _ ) is the energy of the state _ x _ and _ b_(_x _ ) is a trial probability distribution . simplifying the equation ( 4 ) , we draw up the following equation .
@xmath29 by setting _ t _ to 1 in equation ( 1 ) , we can obtain : @xmath30 then we take the equation ( 6 ) into ( 5 ) and acquire : @xmath31 since _
b_(_x _ ) is a trial probability distribution , the sum of the probability distribution should be 1 , i.e. @xmath32 . then the equation ( 7 ) can be expressed as the follows .
@xmath33 by analyzing the equation ( 8) , we know that the second term is equal to zero if
_ b_(_x _ )
is equal to _ p_(_x _ ) .
b_(_x _ ) is equal to _ p_(_x _ ) , the partition function can be written as @xmath34 then by taking the equation ( 4 ) into the above equation , we obtain @xmath35 for a factor graph with no cycles , _ p_(_x _ ) can be easily expressed in terms of the marginal probabilities of variables and constraints as the follows .
@xmath36 where @xmath37 is the number of times that the variable @xmath38 occurs in the constraints , _ m _ and _
n _ are the number of constraints and variables respectively , @xmath39 and @xmath40 are the marginal probabilities of constraints and variables respectively . in addition , by analyzing the two partition functions presented in equations ( 1 ) and ( 2 ) , we can see that _
p_(_x _ ) and _ z _ are equal when _
t _ is set to 1 .
thus , we can obtain the following equation from equation ( 1 ) and equation ( 2 ) on account of the equivalents _ z _ and _ p_(_x _ ) .
@xmath41 then the partition function can be expressed as the follows by plugging the equation ( 11 ) and equation ( 12 ) into equation ( 10 ) .
@xmath42 in equation ( 13 ) , when the variable configuration _ x _ is a solution to a csp @xmath0 , the function @xmath43 is assigned 1 , which means that the term @xmath44 evaluates to 0 .
on the other hand , when the variable configuration _ x _ is not a solution to a csp @xmath0 , the term @xmath45 evaluates to 0 .
therefore , whether or not the variable configuration _ x _ is a solution to a csp @xmath0 , the first term in the exponential function must evaluate to 0 . then equation ( 13 )
can be expressed as the follows .
@xmath46 from the above equation , we can learn that the number of solutions of a given csp @xmath0 can be calculated according to the partition function if the marginal probability of variable @xmath47 and the marginal probability of constraint @xmath39 can be obtained . in the following
, we will present an approach to compute the marginal probabilities . in this subsection
, we address a method bp to calculate the marginal probabilities . the belief propagation , bp for short ,
is a message passing procedure , which is a method for computing marginal probabilities [ 18 ] .
the bp procedure obtains exact marginal probabilities if the factor graph of the given csp @xmath0 has no cycles , and it can still empirically provide good approximate results even when the corresponding factor graph does have cycles . to describe the bp procedure , we first introduce messages between function nodes and their neighboring variable nodes and vice versa .
the message @xmath48 passed from a function node @xmath14 to one of its neighboring variable nodes _ i _ can be interpreted as the probability of constraint @xmath14 being satisfied if the variable @xmath4 takes the value @xmath49 ; while the message @xmath50 passed from a variable node _
i _ to one of its neighboring function nodes @xmath14 can be interpreted as the probability that the variable @xmath4 takes the special value @xmath49 in the absence of constraint @xmath14 .
next we concentrate on presenting the details of the bp procedure ( see fig .
at first , the message @xmath51 $ ] is initialized for every edge ( @xmath14 , _ i _ ) and every value @xmath49 .
then the messages are updated with the following equations .
@xmath52 @xmath53 where @xmath54 is a normalization constant ensuring that @xmath50 is a probability , and @xmath43 is a characteristic function taking the value 1 if the variable configuration @xmath17 satisfies the constraint @xmath14 , taking the value 0 otherwise .
the bp procedure runs the equations ( 15 ) and ( 16 ) iteratively until the message @xmath48 converges for every edge ( @xmath14 , _ i _ ) and every value @xmath49 .
when they have converged , we can then calculate the marginal probabilities of each variable and each constraint in the following equations .
@xmath55 @xmath56 where _ c_ and _ c ` " ` _ are normalization constants ensuring that @xmath57 and @xmath39 are probabilities , and @xmath43 is a characteristic function taking the value 1 if the variable configuration @xmath17 satisfies the constraint @xmath14 , taking the value 0 otherwise . as a whole
, the bp procedure organizes a computation that makes the marginal probabilities computing tractable and eventually returns the marginal probabilities of each variable and each constraint which can be used in the partition function . as we know , the bp procedure can present exact marginal probabilities if the factor graph of the given csp @xmath0 has no cycle . and from the whole derivation of the partition function , we understand that all equations address exact results if the factor graph is a tree .
thus , we obtain the following theorem .
the method pfcount provides an exact number of solutions for a csp @xmath0 if the factor graph of the given csp @xmath0 has no cycle .
the above theorem illustrates that pfcount can present an exact number of solutions if the corresponding factor graph of the given csp @xmath0 has no cycle . in addition , even when the factor graph does have cycles , our method still empirically presents good approximate number of solutions for csp .
in this section , we perform two experiments on a cluster of 2.4 ghz intel xeon machines with 2 gb memory running linux centos 5.4 .
the purpose of the first experiment is to demonstrate the performance of our method on random instances ; the second experiment is to compare our method with two other methods on structural instances .
our # csp solver is implemented in c++ , which we also call pfcount . for each instance , the run - time is in seconds and the timeout limit is 7200s . in this subsection , we conduct experiments on csp benchmarks of model rb , which can provide a framework for generating asymptotically hard instances so as to give a challenge for experimental evaluation of the # csp solvers [ 19 ] .
the benchmarks of model rb is determined by parameters ( _ k _ , _ n _ , @xmath58 , _ r _ , _ p _ ) , where _ k _ denotes the arity of each constraint ; _ n _ denotes the number of variables ; @xmath58 determines the domain size @xmath59 of each variable ; _ r _ determines the number @xmath60 of constraints ; _ p _ determines the number @xmath61 of disallowed tuples of each relation .
table 1 illustrates the comparison of pfcount with state - of - the - art exact # sat solvers sharpsat , c2d , cachet , and approximate solvers approxcount , bpcount on csp benchmarks of model rb close to the phase transition point . in the experiment
, we choose random instances @xmath62 ( 2 , _
n _ , 0.8 , 3 , _ p _ ) with _ n
_ @xmath63 \{10 , 20 , 30 , 40}. moreover , since the theoretical phase transition point @xmath64 0.23 for _ k _ = 2 , @xmath58 = 0.8 , _ r _ = 3 , _
@xmath63 \{10 , 20 , 30 , 40 } , we set _ p _ = 0.20 . in table 1
, the instance frb__a__-_b_-_c _ represents the instance containing _ a _ variables , owning a domain with _
b _ values for each variable , and indexing _ c _ ; _ exacts _ represents the exact number of solutions of each instance ; _ apps _ represents the approximate number of solutions of each instance .
note that the exact number of solutions of each instance is obtained by the exact # sat solvers and the csp instances solved by these # sat solvers are translated into sat instances using the direct encoding method .
the results reported in table 1 suggest that the effectiveness of pfcount is much more evident especially for larger instances .
for example , the efficiency of solving instances with 30 variables has been raised at least 74 times ( instance frb30 - 15 - 1 ) . and the instances with 40 variables can be solved by pfcount in a few seconds . furthermore , pfcount presents a good estimate to the number of solutions for csp benchmarks of model rb .
even if the instances scales are relatively large , the estimates are found to be over 63.129% correct except the instance frb20 - 11 - 3 . therefore , this experiment shows that pfcount is quite competitive compared with the other counters .
table 2 presents the performance of our counter pfcount on hard instances based on model rb .
these instances provide a challenge for experimental evaluation of the csp solvers . in the 1st international csp solver competition in 2005 ,
these instances ca nt be solved by all the participating csp solvers in 10 minutes . in the recent csp solver competition
, only one solver can solve the instances frb50 - 23 - 1 , frb50 - 23 - 4 , frb50 - 23 - 5 , frb53 - 24 - 5 , frb56 - 25 - 5 , and frb59 - 26 - 5 ; only two solvers can solve the instance frb53 - 24 - 3 ; four solvers can solve the instance frb53 - 24 - 1 ; and the rest of instances still ca nt be solved in 10 minutes . in table 2 , _ apps _ represents the approximate number of solutions of each instance ; _ expecteds _ represents the expected number of solutions of each instance , which can be calculated as the following according to the definition of model rb [ 19 ] .
@xmath65 where _
n _ denotes the number of variables ; @xmath58 determines the domain size @xmath66 of each variable ; _ r _ determines the number @xmath67 of constraints ; _ p _ determines the number @xmath61 of disallowed tuples of each relation . when _
n _ tends to infinite , _ expecteds _ is the number of solutions of the instances based on model rb . empirically ,
when _ n _ is not very large , _ expecteds _ and the exact number of solutions are in the same order of magnitude .
therefore , _ expecteds _
precisely estimates the number of solutions of the instances based on model rb . by analyzing the results in table 2 , we can see that pfcount efficiently estimates the number of solutions of these hard csp instances .
it should be pointed out that our pfcount is only capable of estimating the numbers of solutions rather enumerating the solutions . in this subsection
, we carry out experiments on graph coloring instances from the dimacs benchmark set .
the # csp solvers compared with pfcount are ilog solver 6.3 [ 20 ] and csp+xors .
table 3 illustrates the results of the comparison of the # csp solvers on graph coloring instances . in this table , _ apps _ is the approximate number of solutions of each instance ; _ exacts _ is the exact number of solutions of each instance calculated by the exact counters .
note that the results presented by the ilog solver and csp+xors are based on [ 16 ] for lack of the binary codes . as can be seen from table 3
, pfcount does nt give good estimate on these structural instances in contrast with the random instances .
however , the run - time of pfcount clearly outperforms other # csp solvers greatly .
this paper addresses a new approximate method for counting the number of solutions for constraint satisfaction problem .
it first obtains the marginal probabilities of each variable and constraint by the belief propagation approach , and then computes the number of the solutions of a given csp formula according to a partition function , which obtained by introducing the free energy and capturing the relationship between the probabilities of the variables and the constraints .
the experimental results also show that the effectiveness of our method is much more evident especially for larger instances close to the phase transition point .
sang , t. , bacchus , f. , beame , p. , kautz , h. a. , and pitassi , t. : combining component caching and clause learning for effective model counting.in : 7th theory and applications of satisfiability testing .
springer , heidelberg ( 2004 ) gomes , c. p. , hoffmann , j. , sabharwal , a. , and selman , b. : from sampling to model counting . in : 20th international joint conference on artificial intelligence , pp .
springer , heidelberg ( 2007 ) kroc , l. , sabharwal , a. , and selman , b. : leveraging belief propagation , backtrack search , and statistics for model counting . in : 5th integration of ai and or techniques in contraint programming for combinatorial optimzation problems , pp
. 127141 .
springer , heidelberg ( 2008 ) angelsmark , o. , jonsson , p. , linusson , s. , and thapper j. : determining the number of solutions to binary csp instances . in : 8th principles and practice of constraint programming , pp .
springer , heidelberg ( 2002 ) bulatov , a. , dalmau , v. : towards a dichotomy theorem for the counting constraint satisfaction problem . in : 44th annual ieee symposium on foundations of computer science , pp .
ieee computer society , los alamitos , ca ( 2003 ) gomes , c. p. , hoeve , w. j. , sabharwal , a. , and selman , b. : counting csp solutions using generalized xor constraints . in : national conference on artificial intelligence , pp .
aaai press , u.s . | we propose a new approximate method for counting the number of the solutions for constraint satisfaction problem ( csp ) .
the method derives from the partition function based on introducing the free energy and capturing the relationship of probabilities of variables and constraints , which requires the marginal probabilities .
it firstly obtains the marginal probabilities using the belief propagation , and then computes the number of solutions according to the partition function .
this allows us to directly plug the marginal probabilities into the partition function and efficiently count the number of solutions for csp .
the experimental results show that our method can solve both random problems and structural problems efficiently .
function ; # csp ; belief propagation ; marginal probability . |
several authors have proposed that the ground state of matter is in the form of bulk quark matter made of the three flavors , up , down and strange .
we adopt the term strange matter ( sm ) to refer to this bulk quark matter .
conversion of nuclear matter to sm is suppressed at ordinary nuclear densities . at very large nuclear densities like those in neutron stars , where the fermi energy is higher than the strange quark mass , the conversion however is possible .
this has led to conjecture that _ strange stars _ that are predominantly made of sm may form by conversion of neutrons in dense neutron stars @xcite .
the conversion is assumed to be triggered at the center of the neutron star ( where the density is highest ) either by presence of a seed sm or by spontaneous @xmath0 -hyperons .
strange stars fall in a mass range similar to that of neutron stars , but they have smaller radii@xcite .
there is mounting , albeit controversial , observational evidence for the existence of strange stars .
likely candidates are the the compact objects 3c58 and rx j1856.5 - 3754@xcite .
the dynamics of transition into sm has not yet been examined in details . according to @xcite ,
the transition is realized by outward diffusion of strange quarks .
the increased fraction of sm as a result of diffusion triggers further conversion into sm . in this case
the conversion may be described as slow combustion at speeds of about @xmath1 .
however , stability analysis in the presence of gravity has shown that this combustion process is unstable to small perturbations @xcite .
once these perturbation grow they increase the effective area available for diffusion .
therefore the conversion may actually propagate in the form of detonation sustained by the enhanced `` burning '' . in the simplest variant ,
the detonation wave ( dw ) is propagating into a neutron matter nm with constant density .
newtonian gravity was used in @xcite to determine the parameters of the resultant strange star and to estimate the energy released in the process .
more realistic models for the propagation of dws should take into account a varying density of nm .
further , next to black holes , neutron stars have the largest gravitational fields in nature .
moreover , the characteristic time of gravitational processes @xmath2 is comparable to the hydrodynamical time scale @xmath3s@xcite .
therefore , it is prudent to consider detonation in the context of general relativistic hydrodynamics . in this paper
we study self - similar models of dw propagation in non - uniform nm , taking into account a self - consistent treatment of gravity according to general relativity ( gr ) .
we consider detonation in static and collapsing backgrounds of nm .
the paper is organized as follow . in
2 we write the equation of state for quark matter . in
3.1 we present the equations of special relativistic hydrodynamics in the absence of gravitation .
self - similar solutions in this case are presented in 3.2 for uniform and non - uniform backgrounds of nm .
we turn to the general relativistic description in 4 where we derive the equations for general relativistic hydrodynamics and present the corresponding numerical self - similar solutions for static and collapsing non - uniform background . in
5 we conclude with a discussion of the results and their implications .
the equation of state ( eos ) for hot ( quark ) matter has the form of simple relation @xcite , @xmath4 we use throughout @xmath5 and @xmath6 to denote the energy density and the pressure , respectively .
this simple eos may be obtained straightforward from the next eos @xcite @xmath7 @xmath8 @xmath9 is the temperature , @xmath10 is the baryon number density , @xmath11 is the chemical potential of quark matter , @xmath12 is the mit bag constant .
the sound speed in such matter is defined as @xmath13 , where the derivation occurs at constant entropy .
it is easy to see that the sound speed remains constant @xmath14 .
the eos [ eos3 ] was used by witten for calculation of the structure of strange stars @xcite .
we shall adopt this simple approximation for modelling of detonation transition of neutron stars into strange one .
first , we describe the propagation of detonation without taking into account the effects of gravity . this will be useful for comparison of our results with the existing literature and will also facilitate the interpretation of the general relativistic solutions to be presented in [ gr ] .
the neglect of gravity is justified if the detonation time scale is smaller than the dynamical time scale .
gravitational effects become important after the end of detonation , but in this section we will only be concerned with describing the detonation phase .
here we describe the relevant equations for detonation in special relativistic hydrodynamics ( srh ) .
we assume that the pressure of nm is negligible compared to its energy density , @xmath15 otherwise there will not be any significant gain in pressure in the transition from nm to sm .
we will also assume throughout the energy density of the sm @xmath16 behind the front is substantially larger than the vacuum energy
@xmath17 which is close to nuclear energy density of usual matter , @xmath18 under the conditions([epn])-([eps ] ) we express these eoss as @xmath19 where @xmath20 and @xmath21 correspond to hot relativistic matter and cold neutron matter , respectively .
values other than @xmath21 can be used to describe approximately the eos of the nm , in these cases @xmath22 can be regarded representing the square of an `` effective '' sound speed for the nm
. we will base our description of detonation in srh on an analogy with the classical theory of detonation @xcite .
therefore , instead of using covariant notation we express the relevant equations in terms of the radial component of the 3-velocity @xmath23 , where @xmath24 is a radial coordinate and @xmath25 is the time measured in units of lengths .
we use the notation @xmath26 for the `` enthalpy ''
. the relevant equations of srh are , @xmath27 where @xmath28 .
+ the first of these equations is the euler equation and the second is the continuity equation for energy density .
these equations determine the evolution of the system in which a detonation front propagates outward from the center of the neutron star converting its into sm . in a frame of reference moving with the front ,
a requirement of the energy - momentum flow be conserved yields the following jump conditions at the front @xmath29 in addition , there is baryon flux conservation @xmath22 @xmath30 here the indexes @xmath31 designate the quantities concerning to nm and to sm , respectively .
so , @xmath32 is the velocity of inflow " of neutron matter in front of dw and @xmath33 is the departure velocity of sm from the front .
if the 3-velocity @xmath34 in ( [ cond ] ) is measured in the local frame of detonation front then , according to the relativistic equation of velocity transformation , in the laboratory local frame it takes the value @xmath35 where @xmath36 is the speed of dw in this frame .
the classical analysis of detonation conditions leads to definition of a normal detonation condition when the speed of hot fluid corresponds to sound speed @xmath37 @xcite . in the case of nm at rest @xmath38 and substituting ( [ en ] ) in the jump conditions ( [ cond ] ) yields
@xmath39 let the detonation wave propagate in cold nm background ( @xmath40 according to([epn ] ) ) , and let the sm behind the front be hot and relativistic ( see ( [ eps ] ) ) .
we will obtain self - similar solutions of ( [ eorh ] ) for this situation . by analogy to classical detonation
, we find the self - similar solutions assuming the velocity @xmath41 to depend on @xmath24 and @xmath42 only through the variable @xmath43 analysis of the system ( [ eorh ] ) shows that self - similar solutions can be obtained in the case of power dependence of density of nm , @xmath44 this distribution of nm can be created by its own gravity ( newtonian gravitation ) .
but for a fast detonation wave ( see ( [ dw ] ) ) , the inhomogeneity of nm distribution is only important .
( similar situation is realized for a strong shock wave in an inhomogeneous atmosphere @xcite . )
according to ( [ dw ] ) , we have at the front position @xmath45 , @xmath46 and the front velocity remains constant .
therefore , the density of sm at the front varies with time according to the relation : @xmath47 this relation motivates the following form for @xmath41 and @xmath48 , @xmath49 by expressing the partial derivatives @xmath50 and @xmath51 in terms of @xmath52 the eqs .
( [ eorh ] ) are transformed into the following ordinary differential equations , @xmath53 } } \mathord{\left/ { \vphantom { { { \left [ { { \frac{{2v\gw ^{2}}}{{\xi ( 1 - v\xi ) } } } - { \frac{{3}}{{4}}}{\frac{{k\gw ^{4}}}{{(1 - v\xi ) ^{2 } } } } } \right ] } } { { \left [ { { \frac{{1}}{{c_{s}^{2 } } } } \left ( { { \frac{{v - \xi } } { { 1 - v\xi } } } } \right)^{2 } - 1 } \right ] } } } } \right . \kern-\nulldelimiterspace } { { \left [ { { \frac{{1}}{{c_{s}^{2 } } } } \left ( { { \frac{{v - \xi } } { { 1 - v\xi } } } } \right)^{2 } - 1 } \right]}}},\ ] ] @xmath54 we first present solutions for the propagation of dw in a homogeneous distribution of nm ( @xmath55 ) . in the case
the equations eqs.([eorh2_1])-([eorh2_2 ] ) become equations is @xmath56 } } \mathord{\left/ { \vphantom { { { \left [ { { \frac{{2v\gw ^{2}}}{{\xi ( 1 - v\xi ) } } } } \right ] } } { { \left [ { { \frac{{1}}{{c_{s}^{2 } } } } \left ( { { \frac{{v - \xi } } { { 1 - v\xi } } } } \right)^{2 } - 1 } \right ] } } } } \right . \kern-\nulldelimiterspace } { { \left [ { { \frac{{1}}{{c_{s}^{2 } } } } \left ( { { \frac{{v - \xi } } { { 1 - v\xi } } } } \right)^{2 } - 1 } \right ] } } } ; \quad { \frac{{d\ln \varepsilon } } { { d\xi } } } = { \frac{{4}}{{{\gw ^{2}}}}}\left({{\frac{{\xi - v}}{1 - v\xi } } } \right){\frac{{dv}}{{d\xi } } } , \ ] ] and in the non - relativistic case ( @xmath57 ) the equation for @xmath41 reduces to zeldovich s classical equation for detonation @xcite , @xmath58 ) for propagation of detonation wave in a homogeneous ( a ) and an inhomogeneous background ( b , c ) . in ( a ) there is a solution for velocity @xmath59 of gas that leaves detonation front propagating with velocity @xmath36 and comes to node of solutions @xmath60 , where it may be matched to the solution for rest matter @xmath61 . for @xmath62 ( panel b
) node shifts in the plan @xmath63 .
for @xmath64 ( panel c ) there in no node but we have a unique solution leaving detonation front and converging to @xmath61,height=6 ] analysis of the first equation of ( [ eorh1 ] ) in the phase plane @xmath65 reveals a variety of solutions which all converge to a node at @xmath66 and @xmath61 with the derivative @xmath67 .
the solutions can be obtained independently for the ranges @xmath68 and the @xmath69 .
the solutions can be found by numerical integration of the equation ( [ eorh1 ] ) starting from an arbitrary point in phase space .
examples of these solutions are shown in figure ( [ sr1 ] , a ) . as is clear from this figure the only physically accepted solution in the range @xmath70 is @xmath71 ( see @xcite ) shown as the thick solid line . to the right of @xmath72
the solution shown as the dashed dotted line corresponds to normal detonation . for this solution the velocity at the front @xmath73 is equal to the sound speed ( see ( [ dw ] ) ) and
is zero at @xmath72 .
the derivative of the velocity diverges as @xmath74 approaches @xmath36 , and tends to zero at @xmath75 .
this detonation solution describes hot matter at rest for @xmath76 and moves outward for @xmath77 .
the velocity of the front is @xmath78 , according to ( [ dw ] ) .
our solution for the propagation of detonation in a uniform distribution of nm is fully consistent with previous calculations in the literature @xcite .
we also refer the reader to the literature treating the similar problem for relativistic detonation in `` scalar field '' stars @xcite .
we search now for self - similar solutions of eq .
( [ eorh2_1 ] ) for @xmath79 , i.e. , for the propagation of detonation in nm with decreasing density . for @xmath79 , also , the equation has a node to which all solutions converge with the identical derivative . in this case , however , the node is at @xmath80 and not at @xmath61 as in the @xmath55 case .
examples of the solutions are shown in panels b and c of fig.1 .
as in the homogeneous case , the physical solution corresponds to with @xmath81 at @xmath82 .
the point @xmath83 is a critical one ( saddle ) and the solution originating from this point is the separatrix and is shown as solid line in the figure .
near @xmath84 we find that @xmath85 for @xmath86 .
for @xmath87 the node lies to the left of dw wave at @xmath88 .
therefore , self - similar solutions corresponding to normal detonation ( cf . [ dw ] ) ) exist only for @xmath89 . for @xmath90 the node shifts to the right of the dw .
in this case , there are no self - similar solutions for normal detonation in a static neutron configuration .
but we can find self - similar solutions for over - compressed detonation .
for such type of detonation the velocity of the hot matter at the front is determined by the details of the actual physical situation but it is less than sound speed in sm @xcite the position of dw front is defined from ( [ cond ] ) , @xmath91 , @xmath92 then a laboratory velocity of outgoing sm from the front of wave ( [ vel ] ) will be supersonic nearby the front of dw but transition through the sound speed in the self - similar solution occurs continuously and @xmath93 at @xmath94 according to the formula ( [ separ ] ) .
the conditions ( [ placedf ] ) are actually conditions of so - called `` over - compressed '' detonation . in our statement of the problem , at nm densities
less than @xmath95 @xmath96 we deal with a transition mechanism based on diffusion of strange quarks from sm to nm
. corresponding velocity of diffusion @xmath97 depends on density of nm @xcite , but in principle , it should be less than sound speed in sm .
therefore the velocity of sm leaving the front is equal to the velocity of the anomalous ( turbulent ) diffusion @xmath97 , which may be higher than the estimate given in @xcite ) .
as we shall consider a spherically - symmetric problem , it is convenient to write the metric as @xmath98 here the standard spherical system of coordinates @xmath99 is used .
variables @xmath100 are functions of @xmath101 . to derive self - similar solution in form
similar to srh analysis ( [ eorh ] ) we will use an `` auxiliary '' time @xmath102 instead of the proper time of local observer @xmath42 .
the transformation from variables @xmath103 to new variables @xmath104 is carried out with help of the jacobian@xcite : @xmath105 the system of hydrodynamics equations in gr are then @xmath106 @xmath107 the einstein equations for the metrics can be written down as : @xmath108 where @xmath109 are newtonian gravitation constant and speed of light , respectively and @xmath110 . the equations ( [ eorhgr1_1])-([eogr ] ) can be presented as system of three equations by eliminating of function @xmath111 , @xmath112 in the limit of @xmath113 this system of equations reduces to the srh equations ( [ eorh2_2 ] ) .
in contrast to srh , where self - similar solutions could be found for all @xmath114 , in grh the full system ( [ fulls ] ) permits only self - similar solutions only with nm distribution density ( [ en_dist ] ) with @xmath115 .
we show in appendix [ ap ] that this profile describes approximately a realistic density profile in a neutron star with a harrison - wheeler - wakano equation of state@xcite . in analogy with srh
we make the ansatz , @xmath116 adopting an eos of the form ( [ en ] ) and using @xmath117 expressed in terms of the self - similar variable @xmath118 , the grh equations are expressed as , @xmath119/f_2\end{aligned}\ ] ] @xmath120}/f_2\end{aligned}\ ] ] @xmath121 in the case of hot sm ( @xmath20 ) , eqs .
( [ full1])-([full3 ] ) have a form , @xmath122}/\left[3\left({\frac{v - \xi}{1 - v\xi}}\right)^{2}-1\right],\ ] ] @xmath123 where @xmath124 .
we first investigate the boundary conditions imposed on this system of equations ( [ fd3 ] ) .
we assume that at @xmath125 we have @xmath126 in the vicinity of @xmath82 we solution can be approximate , up to third order in @xmath118 , as @xmath127 where @xmath128 the static solution of the system ( [ full3]-[full1 ] ) for cold nm background @xmath129 is also the analytical solution of the oppenheimer - volkoff equation ( see appendix)in the case @xmath130 .
this full solution of [ fd3 ] for hot quark matter is found numerically and shown as the solid line in fig.[fig3 ] .
it corresponds to the separatrix by analogy to the same problem in srh ( shown as the dash - dot line in the figure ) and is valid mathematically up to some @xmath131 .
however , the solution must be terminated at the dw position , @xmath36 . in the case of normal detonation ,
the the position @xmath36 on the axis @xmath118 is found by the requirement that the speed of outgoing sm @xmath34 is equal to the sound speed .
the previous analysis has shown , however , that for dw propagating in a background @xmath132 such point can not exist without taking into account gravitation .
the gravitational influence results in backward motion of node towards the front of normal detonation in the phase plane but jump conditions of normal detonation exist only in hot neutron stars with @xmath133 .
this dynamical process may convert hot ns into ss without any appreciable thermal energy release . in cold ns
we must deal with over - compressed detonation ( see section [ isr ] ) .
a numerical calculation shows that @xmath34 can be a little less than @xmath134 . for subsequent analysis of numerical models of transition we will be using analytical approximation of neutron star ( see [ ap ] ) .
it satisfactorily describes distributions of density @xmath135 , pressure @xmath136 and the metrics @xmath137 of neutron star from some initial dimensionless radius @xmath138 .
thus @xmath139 according to ( [ placedf ] ) , the pressure of sm matter behind the dw front with account of ( [ approx2 ] ) will be defined by relation : @xmath140 since the front of wave propagates with constant speed @xmath36 and time @xmath102 , as well as @xmath24 , is measured in units of @xmath141 and @xmath142 .
the initial moment of time is chosen so that the front of wave is found at some @xmath138 where the analytical approximation of distribution of ns density is true .
( note , that true distribution has a maximum in the center and a singular analytical solution approximates only the decreasing of this density with growth of @xmath143 starting with finite @xmath144 .
it is supposed that the specified self - similar solution realizes from the moment of achievement of point @xmath138 by the dw front and further . )
thus for the representative of pressure at the front of wave we have @xmath145 and position of @xmath36 is defined in the following way : at the point @xmath73 values @xmath118 , @xmath41 and @xmath34 have to obey the law of energy - momentum flow conservation ( [ cond ] ) and the relativistic equation of velocity transformation simultaneously : @xmath146 thus , we are starting the numerical solution from point @xmath147 taking into consideration an analytical research of the solution in the vicinity of @xmath148 ( [ init ] ) . at that , the setting of @xmath149 and @xmath150 should be chosen so that after definition of the front position they would be corresponding to the mentioned values ( [ pres ] ) and ( [ mtr ] ) .
recent observational data strongly suggest that `` long - duration '' gamma ray bursts ( grbs ) are associated with supernova explosions@xcite .
these grbs follow the supernova events after a period varying from hours to months .
this may point to the possibility that these grbs appear as a result of the conversion of nm into sm during collapse of the sn remnant .
this occurs when the mass of the nm becomes larger than the chandrasekhar mass limit .
the detonation in this case propagates on the background of the falling nm , so normal detonation wave forms .
negligible pressure forces in the collapsing nm allow us to neglect pressure in this statement of the problem .
there are only two relevant parameters determining the self - similar solutions .
these are the mass and the radius of the remnant strange star .
the parameters of obtained self - similar solution are compare with those of the observational data . in this statement of problem
the jump conditions on the front ( [ cond ] ) yield @xmath151 as in the stationary case , we are solving system ( [ fd3 ] ) for hot strange matter from point @xmath152 to place of the front @xmath153 the solutions of system ( [ full1])-([full3 ] ) for nm and sm are presented in fig.([col ] ) . at that we will claim the metric @xmath154 to be defined by mass @xmath155 of the strange star at the radius @xmath156 ( at @xmath157 ) . here
we use parameters @xmath158 and @xmath159 ( see data in @xcite ) . as an example of detonation in static neutron star ,
let us consider the model of detonation transition of neutron star with mass of central part @xmath160 and radius @xmath161 measured in characteristic lengths @xmath162 that corresponds to @xmath163 .
the analytical approximation of density distribution @xmath164 measured in units of @xmath165 is set by expression ( [ delt ] ) with @xmath166 . at that @xmath167 .
the results of numerical research for the over - compressed detonation are submitted in fig .
[ fig3 ] . in the example under consideration
the velocity of dw is equal @xmath1680.85 , that is close to the velocity @xmath169 .
it is interesting to estimate the value energy concerned with propagation of hot matter behind dw front , when the last one reaches the specified boundary radius of neutron configuration @xmath170 : @xmath171 } } } dr = 16\pi { \int\limits_{0}^{r } { { \frac{{v^{2}}}{{\gw ^{2}}}}p\,r^{2 } } } dr\ ] ] whence , using ( [ rel ] ) and ( [ pres ] ) , we shall have @xmath172 the value of the dimensionless integral , obtained on the basis of the numerical solution , gives the value @xmath173 , corresponding to @xmath174 .
this is close to the the energy produced in a gamma - ray burst @xcite . the gravitational mass defect @xmath175 .
according to @xcite we have , @xmath176 or , using ( [ rel ] ) and ( [ pres ] ) , @xmath177 thus , we have @xmath178 .
the energy release associated with detonation can be roughly estimated as @xmath179 . in our model
it is only about @xmath180 of the mass @xmath155 .
the detonation will stop when the nm density is close to @xmath181 , leaving a crust made of the original matter of the neutron star .
this crust can be expelled if the kinetic energy of the sm is large enough .
the kinetic energy needed for that can be estimated by comparing the binding energy of the crust , @xmath182 to the energy release , @xmath183 .
we can see that in the case @xmath184 , @xmath185 which is less than the energy release @xmath186 .
therefore for this value @xmath22 the crust might be expelled , leaving a bare strange star .
table 1 lists estimates for other values of @xmath22 and for collapsing nm .
ccccccccc + @xmath187 & n & @xmath188 & @xmath156 & @xmath189&@xmath190 & @xmath191 & @xmath192 + 9.40&0.2&0.59 & 5.23&-0.51&35.18 & 38.49&20.94 + 6.71&0.16 & 0.60 & 5.42 & - 0.52 & 50.00&21.35 & 22.13 + 5.37&0.08 & 0.573&8.37&-0.52&48.44 & 13.24&28.27 + collapse&0&1.1&9.74&-1/@xmath193 & 87.72 & 58.64&- +
we have studied the conversion of neutron matter into strange matter as a detonation wave propagating outward of a central region .
the dynamics of neutron - quark transition has been studied earlier for a uniform static background of neutron matter in the context of special relativity @xcite . here
we employ special and general relativistic hydrodynamics to find self - similar solutions for the propagation of detonation in a non - uniform background of neutron matter .
both static and collapsing backgrounds are considered .
solutions for detonation in a static background may correspond to the conversion of neutrons in an otherwise stable neutron star . in a collapsing background
the solutions may describe conversion in the neutron core of sn type ii @xcite . in this latter case detonation
is assumed to be triggered at the onset of the collapse .
a description of the physical processes leading to detonation can be found in papers @xcite .
briefly , the initial stages of conversion of neutrons to strange quark matter proceeds by diffusion of sm into the neutron matter .
the diffusing sm then stimulates further conversion and the whole process may be described as combustion @xcite .
however , the process turned out to be hydrodynamically unstable when gravity is taken into account @xcite .
hydrodynamical instabilities increase the surface area available for diffusion allowing for the possibility of propagation in the form of detonation . in the other words ,
these processes leads to anomalous diffusion of strange quarks to the detonation front with speed that is equal to speed of leaving hot matter . in special relativity
the self - similar solutions describe detonation in a static background of nm with distribution @xmath194 .
it was shown the condition of a normal detonation exist for @xmath87 , while denotation must be over - compressed for @xmath195 .
self - similar solutions of the problem in general relativity are permitted only for @xmath64 .
the reason for this restriction is the appearance of new fundamental constant @xmath196 in gr .
this additional physical scale prevents the development of a self - similar behavior of the type considered here . in special relativity
the detonation for @xmath64 is over - compressed independently of the equation of state of cold matter .
however , in general relativity we obtain normal detonation for @xmath22 close to @xmath197 ( @xmath198 ) . below a limiting value of @xmath22 only over - compressed detonation
is possible .
we have not been able to find the limiting value of @xmath22 analytically but our numerical calculations show that it should be @xmath199 .
normal detonation is obtained for conversion in collapsing cold nm . in this case the pressure of nm is completely neglected , i.e. , @xmath21 in ( [ en ] ) . in such statement of problem one can easily calculate all relevant quantities using only two parameters characterizing the system , e.g. , the mass and radius of the strange star .
the propagation of detonation in neutron stars must halt at the radius where the neutron density is close to the quark vacuum energy , @xmath200 .
when the dw reaches this radius the sm would be surrounded by a neutron crust .
the description in this paper breaks down at this stage of the evolution .
however , we still can use the self - similar solutions to determine the energetics of the system and to predict some general features its subsequent evolution . in all of our self - similar solutions , except those describing detonation in a uniform background , the velocity of sm is moving outward .
although the `` kinetic energy '' of the sm is insufficient to disassemble the whole configuration ( see table 1 ) , it might be enough to expel the crust surrounding the quark core . according to table 1 ,
if @xmath22 is close to @xmath197 , the crust is maintained . in this case
we may end up with a crusted strange star . for lower @xmath22
the energy is sufficient to remove the crust , leaving a a bare strange star . in our estimates
presented in table 1 we have neglected a few factors .
non - vanishing mass of strange quark and constant of strong interaction taken into account will lead to reducing of energy release .
we also have neglected any further conversion ( e.g. , by slow combustion in the crust ) after the detonation is halted .
we wish to thank a.starobinsky for useful discussion of the model of relativistic detonation which have been considered by v.g . and v.f . in @xcite .
also we acknowledge to v.usov for affirmative remarks that they made . 99 a.r.bodmer,_phys .
rev._*d4 * , 16012 ( 1971 ) e. witten , _ phys .
rev . _ * d30 * , 272 ( 1984 ) .
alcock c. , farhi e. and olinto a. , 1986 , apj * 310 * , 261 p.o.slane,d.j .
helfand , nad s. s. murray , 2002 , apj * 571*,l45 j. j.drake , h. l.marshall , s.dreizler , p. e.freeman , a.fruscione , m.juda , v.kashyap , f.nicastro , d. o.pease , b. j. wargelin , k. werner , 2002 , apj * 572 * , 996 a. olinto , _ phys .
lett . _ * b192 * , 71 ( 1987 ) .
horvath j.e . and benvenuto o.g
. , 1988 , phys .
bf b 213 , 516 .
benvenuto , j.e . horvath and h. vucetich ,
_ int . journ . of mod .
* 4 * , 257 ( 1989 ) .
l. landau and e. lifshitz , _ fluid mechanics _ ( pergamon , oxford,1987 ) .
y. raizer and ya .
zeldovich , _ physics of shock waves and high - temperature hydrodynamic phenomena _ ( academic press , new york , 1967 ) .
v. folomeev , v. gurovich and r. usupov , _ phenomen of gamma - ray bursts as relativistic detonation of scalar fields _ , astro - ph/0008334 ; v. folomeev , v. gurovich , h. kleinert , h.j .
schmidt , _ grav.cosmol .
_ * 8 * , 299 ( 2002 ) ; v.ts .
gurovich , v.n .
folomeev , _ jetp lett . _ * 76 * , 604 ( 2002 ) .
stanyukovich , o. sharshekeev and v.ts . gurovich _
soviet physics - doklady _ * 10 , no .
11 * , 1030 ( 1966 ) .
harrison , k.s .
thorne , m. wakano and j.a .
wheeler , _ gravitation theory and gravitational collapse _
( the university of chicago press , chicago ) .
i. bombaci , i parenti and i. vidana , astro - ph/0402404 t. piran , astro - ph/9907392 .
s.l . shapiro and s.a .
teulkolsky , _ black holes , white dwarfs , and neutron stars - the physics of compact objects _
( wiley - interscience , new york ) .
lugones g. , benvenuto o.g . and vucetich h. , 1994 , phys
* d 50 * , 6100
. l. landau and e. lifshitz , _ the classical theory of fields _ ( pergamon , oxford , 1975 ) .
in this section we solve the oppenheimer - volkoff equation to compute the density profile in neutron stars with the harrison - wheeler - wakano ( hww ) equation of state @xmath201 where @xmath202 .
the hww eos is valid for central densities @xmath203 @xcite .
the sound speed @xmath204 is given by @xmath205 therefore for small @xmath135 this eos describes cold nm . from equations ( [ eogr ] ) , after integration
, we have @xmath206 where @xmath207 is the energy density of nm and @xmath208 is the mass of sphere within a radius @xmath24 .
the oppenheimer - volkoff for the equilibrium of ns is @xcite @xmath209 } } { { rc^{2}[r - 2(g / c^{2})m(r)]}}}.\ ] ] to present eqs .
( [ gd1])-([op - volk ] ) in dimensionless form , we use the radial coordinate @xmath143 defined as @xmath210 where the characteristic length @xmath211 . in dimensionless units the mass within a radius @xmath24 is @xmath212 in order to get the physical mass we multiply @xmath11 by @xmath213 oppenheimer - volkoff s equation ( [ op - volk ] ) in dimensionless variables can be presented as @xmath219 this equation together with eq .
( [ mass ] ) and eos ( [ hw ] ) completely defines the structure of the neutron star .
given the density at the center and assuming a vanishing pressure at the star s boundary , a unique solution to these equations can be found by numerical integration .
we denote the density at the center ( @xmath220 ) by @xmath221 .
the numerical solutions are shown as the dashed lines in the three panels in fig.[fig1 ] . also , the solid lines in these panels show solutions obtained by assuming @xmath222 thus we have : @xmath223 it is easy to see that this special solution satisfies to eq .
( [ op - volk_1 ] ) where the coefficient @xmath224 is defined by expression : @xmath225 at that the components of the metric @xmath226 and @xmath227 are given by the formula @xmath228 where @xmath229 ( here @xmath230 is an integration constant ) .
it is interesting to note that the @xmath226 is constant , while @xmath231 depends on @xmath143 .
this solution has a conic singularity at @xmath220 .
but the metric at this point has no singularity .
we approximate density profiles of neutron stars by shifting this solution at @xmath232 .
comparison of this simple analytical solution with a numerical calculation is presented in fig . | we study the conversion of neutron matter into strange matter as a detonation wave . the detonation is assumed to originate from a central region in a spherically symmetric background of neutrons with a varying radial density distribution .
we present self - similar solutions for the propagation of detonation in static and collapsing backgrounds of neutron matter .
the solutions are obtained in the framework of general relativistic hydrodynamics , and are relevant for the possible transition of neutron into strange stars .
conditions for the formation of either bare or crusted strange stars are discussed . |
in order to compare the large - scale characteristics of galaxies objectively , quantitative measures of the structural components are necessary .
there are many schemes for describing the structure of galaxies , including hubble classification ( sandage , 1961 ) , isophotal radii ( holmberg , 1947 ) , concentration parameters ( kent , 1985 ; kodaira et al . ,
1990 ) , and the use of various fitting functions ( devaucouleurs , 1953 ; freeman , 1970 ; kormendy , 1977 ) .
all these techniques specify parameters which can provide insight into the formation and evolution of galaxies .
the use of standardized fitting functions is arguably the most powerful method for measuring the large - scale structure of galaxies , as the functions yield a variety of parameters which can be easily compared with the results of theoretical models .
they also provide a reasonably detailed description of the radial light distribution with a small number of parameters .
ideally , fitting functions would be based upon the physics of the formation and evolutionary processes .
unfortunately , these processes are neither simple nor well - understood , so the most commonly used functions are derived empirically .
traditional fitting functions for elliptical galaxies and spiral galaxy bulges include the hubble law ( hubble , 1930 ) , king model ( king , 1966 ) , and devaucouleurs law ( devaucouleurs , 1953 ) .
recently , there has been some work which suggests that a generalized version of the devaucouleurs profile ( @xmath0 ) provides for better bulge fits ( andredakis , peletier , and balcells , 1995 ) , and that late - type spirals often have bulges which are best fitted by exponentials ( andredakis and sanders , 1994 ) .
exponentials ( freeman , 1970 ) and inner - truncated exponentials ( kormendy , 1977 ) work well for the disk components of spiral galaxies .
overall , the devaucouleurs law seems to be quite effective as a fitting function for bulges ; it can be written as @xmath1 where @xmath2 is the surface intensity of the bulgeat radius @xmath3 , @xmath4 is a characteristic radius defined to be the radius within which half the total light is emitted , and @xmath5 , the effective intensity , is simply the surface intensity at @xmath4 .
( in this paper , we will consistently use the term `` surface intensity , '' @xmath6 , to mean linear intensity units per square arcsecond , and `` surface brightness , '' @xmath7 , to mean the same quantity in magnitude units . ) similarly , the inner - truncated exponential is defined by kormendy ( 1977 ) as @xmath8 where @xmath9 is the disk surface intensity at radius @xmath3 , @xmath10 is the central intensity of the disk , @xmath11 is the disk scale length , and @xmath12 is the radius of the central cutoff of the disk ( `` hole radius '' ) ; the pure exponential disk is the same as equation ( 1 - 2 ) with @xmath13 .
kormendy ( 1977 ) found that a value of @xmath14 in the truncation term works well , and we have adopted @xmath15 for all of the fits presented here .
figure 1 illustrates the usefulness of including a truncation term in the fitting function , using ngc 3145 as an example .
others have used these fitting functions to obtain the relevant structural parameters for spiral galaxies in order to compare galaxies of different types , luminosities , and environments .
for example , boroson ( 1981 ) fit brightness profiles for 26 non - barred spiral galaxies in order to determine how the bulge - to - disk ratios are related to the hubble types , and to investigate the relationship between spiral and s0 ( lenticular ) galaxies .
kent ( 1985 ) performed a similar analysis using 105 intrinsically luminous galaxies of all types .
kormendy ( 1977 ) decomposed the brightness profiles of seven compact s0 galaxies and one `` normal '' galaxy to check freeman s ( 1970 ) hypothesis of a constant central disk surface brightness .
more recently , dejong ( 1996 ) investigated the freeman ( 1970 ) result of a constant central surface brightness of disks , and some other relationships between the fitting parameters and the hubble sequence , using @xmath16 and @xmath17 brightness profiles of 86 face - on disk galaxies .
a relatively recent innovation is to fit a surface to the entire galaxy image ( byun and freeman 1995 ) using the above fitting functions and also solving for the center and ellipticity of the projected distributions .
a general advantage of this approach is that the bulge and disk components can be allowed to have different ellipticities , which alleviates the problem of projection effects for moderate- to high - inclination systems : because the rounder bulge is typically sampled at a smaller galactocentric radius than the inclined disk for a given position in the image , the derived bulge parameters are systematically too large when estimated from brightness profiles obtained by azimuthal averaging techniques .
however , the profiles used here are major axis cuts ( see below ) , so this difficulty should not affect our fitted parameters ( burstein 1979 ) .
the cost of using major axis cuts is that of throwing away much of the information in the images .
all of these programs except kormendy ( 1977 ) employed a simple exponential to describe the disk light distribution .
as part of a study of the origin of inner - truncated spiral galaxy disks , or type ii brightness profiles ( freeman , 1970 ) , we have used the devaucouleurs law and the inner - truncated disk fitting function from kormendy ( 1977 ) for the bulge - disk decomposition of 659 spiral and lenticular galaxy brightness profiles .
our preliminary study ( baggett , baggett , and anderson , 1993 ) indicated that a substantial fraction of all spiral galaxies exhibit an inner - truncation , so the inclusion of such a term in the fitting function seems justified with this large set of brightness profiles .
furthermore , the data set used in this study is extremely homogeneous , all images having been obtained , reduced , and analyzed in the same way .
thus , the results of our fitting should be a useful resource for many studies of the large - scale properties of disk galaxies .
the following sections will describe the data and the bulge - disk decomposition procedures , and will present the fitting results together with a discussion of the associated errors .
the brightness profiles used for this study were obtained from kodaira et al .
( 1990 ; hereafter , panbg ) in a machine - readable form .
the initial panbg sample of galaxies was selected on the basis of being included in _ a revised shapley - ames catalog of bright galaxies _
( sandage and tammann , 1981 ; hereafter , rsa ) and being north of declination -25 . of the 911 such galaxies in the rsa , 791
are included in the panbg , and 659 of those have hubble types ( t - types ) from the _ third reference catalogue of bright galaxies _ ( devaucouleurs et al . , 1991 ;
hereafter rc3 ) in the range -3 to 9 , which indicates that they are spiral or lenticular galaxies .
these 659 galaxies form the basis of our study .
galaxies in the panbg were observed photographically over a period of almost two decades ( late 1970 s through 1988 ) with the kiso observatory 1.05-m schmidt telescope , using kodak iia - d plates and a schott gg495 glass filter to define the `` photographic v - band . ''
exposure times ranged from 30 minutes to 60 minutes , with 50 minutes being standard , and the plates were developed in fuji pandol or kodak d-19 developer .
the plates were then digitized with the kiso observatory pds microdensitometer utilizing a 1 square aperture , except for ngc 224 and ngc 598 , which were measured with a 10 square aperture .
each plate included a step wedge image which was scanned in the same fashion as the galaxy images .
measured densities were converted into relative intensities via the wedge calibrator , and the aperture photometry from longo and devaucouleurs ( 1983 ) was then used to transform the resulting magnitudes to a standard photometric system .
the stated internal photometric accuracy in the panbg is about 0.1 magnitudes ( standard deviation ) , and is dominated by errors in the absolute calibration .
the brightness profiles in the panbg were obtained from the resulting calibrated images by taking a cut along the apparent major axis of each galaxy .
the major axis was defined by fitting the 25 @xmath18 isophote with an ellipse whose center was fixed at the center of gravity of a 21x21 pixel region around the apparent nucleus of the galaxy . the surface brightness was then sampled along this axis using a circular aperture which was stepped outward from the galaxy center in 2-pixel steps ( 2 for all but ngc 224 and ngc 598 , which used 20 steps ) .
the aperture used was 2 pixels in diameter at the galaxy center and the diameter was varied in such a way as to be tangent to a sector with a 5 degree vertex angle centered on the major axis .
this scheme of varying the aperture size was chosen in order to compensate for the decreased signal - to - noise ratio in the outer portions of the galaxies . as a result
, there is some radial smearing of the intensity information at large galactocentric radii , smoothing structure in the outer portions of each profile . further smoothing results from our averaging of the two halves of the major axis cut to produce the final profiles . for full details of the data
acquisition and reduction processes , the reader is referred to panbg .
the major axis brightness profiles were fitted using a combination of a devaucouleurs ( 1953 ) law ( equation 1 - 1 ) and an inner - truncated exponential ( kormendy , 1977 ; equation 1 - 2 ) .
the interactive stsdas task nfit1d was used for all of the fitting ; this task uses the downhill simplex algorithm for performing a non - linear least - squares fit of the data to a specified function , and allows interactive control over the inclusion of the various parameters and the range of the data to be fit .
it is a very flexible routine and we found that it accurately returns the values of the fitting parameters in a number of test cases .
fitting is performed on the surface intensity data and is accomplished by minimizing the weighted rms deviation of the data from the fit .
the most appropriate weighting function , @xmath19 , is one which uses the variance of the intensity measurement for each point as its basis , with the weight of the @xmath20 point being @xmath21 where @xmath22 is the variance of the @xmath20 point ( bevington , 1969 ) .
we chose to use a weighting based on the poisson distribution , where @xmath23 , as this was consistent with the fashion in which the intensity measurements were obtained .
an unfortunate side effect of this @xmath24 weighting function is that it destroys the usefulness of the weighted rms residual as a goodness of fit measure _ between _ galaxies .
the value of the weighted rms residual is highly dependent upon the fitting range , with fits to lower surface intensities being virtually guaranteed a lower weighted rms value than fits stopping at higher surface intensities . as a result , the weighted rms residual is a useful diagnostic only during the fitting process for a given galaxy and , as such , the weighted rms values for each fit are not reported here .
however , the _ unweighted _ rms residuals , @xmath25/n}$ ] , can be computed after the fact , and we have tabulated these values , expressed in magnitude units , as a basis for assessing the relative quality of the fits .
these residuals were computed from the portions of each brightness profile at radii larger than 3 and out to the point where the profile first drops to a surface brightness fainter than 25 @xmath18 .
hence , all of the calculations avoid the portion of the profile most affected by seeing , and reach the same limiting surface brightness .
also , any structure present in the profile contributes to this measure , and thus galaxies with significant structure will be recognizable by their correspondingly larger rms value . in this way ,
the unweighted rms residuals are directly comparable from galaxy to galaxy , and reflect more accurately the quality of the fits than do the weighted rms residuals . aside from the weighting function ,
the only other controlled aspect of the fitting was the choice of components to include in each fit .
generally , all fits were attempted with both a bulge and a pure exponential disk , resulting in the estimation of four quantities and their uncertainties : @xmath26 , @xmath27 , @xmath28 , and @xmath29 ; inner - truncated disks , with the additional parameter , @xmath30 , were only utilized if the profile had the suggestion of a plateau near the center .
if any component ( bulge , disk , inner - truncated disk ) was fit with a negative value for a coefficient , that component was deemed non - physical and was removed from the fitting function . in cases where the need for a specific component was not obvious , fits with and without it were obtained and the component was included if the _ weighted _ rms value was smaller by at least 10% than the fit without the component .
a sample of the results of the fitting are presented in table 1 , which provides for each galaxy its ngc / ic designation ( as indicated by a leading `` n '' or `` i '' , respectively ) , its revised hubble type , t - type , and axis ratio from the rc3 , and the fit parameters determined here .
the fit parameters are given as the fitting range in arcseconds ( in the format minimum : maximum ) , the bulge effective surface brightness ( @xmath31 ) , the bulge effective radius ( @xmath4 ) , the disk central surface brightness ( @xmath32 ) , the disk scale length ( @xmath11 ) , and the disk hole radius ( @xmath12 ) .
the table also includes the seeing as reported in the panbg , and the seeing - corrected values of the fitted bulge parameters ( @xmath33 and @xmath34 ; see the appendix ) .
the last two columns contain the unweighted rms deviation of the fit from the profile in @xmath18 and a column of notes .
all surface brightness quantities are in units of @xmath18 , and all radii are in units of arcseconds .
no corrections for galactic extinction , internal extinction , or inclination have been applied the fits are for the observed major axis profile .
we have chosen to present the results for the observed profiles in order to allow the reader the opportunity to apply the corrections ( s)he deems most appropriate ; we therefore avoid the necessity of uncorrecting our fits and subsequently applying a different correction .
the fitting procedure did not include any allowance for the effects of seeing on the brightness profile . in the appendix
, we describe an experiment which was designed to quantify the impact on the fitted bulge parameters of excluding seeing from the fitting ; the net result is that our fitted @xmath35 tends to be too large by up to an order of magnitude when the seeing is much larger than the input effective radius , and our fitted @xmath36 tend to be slightly too small in the same cases .
this is what is intuitively expected , of course .
although the profile definition procedure introduces a significant level of radial smearing , structure is still apparent in many of the profiles .
the fitting process nominally made no allowance for the presence of structure , fitting across arms and bars as if they were simply noise in the data , except in some specific instances .
these instances occur when a very strong , isolated feature is present then the radial range occupied by the feature was excluded from the fit . if a strong feature is present at the end of the fitting range , the fitting range was suitably shortened .
the existence of such a condition was manually determined , and is indicated in table 1 by the presence of multiple ranges in the fitting range column and/or by a note ( `` a '' ) in the last column of the table .
figure 2 presents the profiles and fits for the galaxies from table 1 .
each plot shows the observed profile from the panbg as crosses , and the fitted bulge and disk components and their sum as solid lines .
in addition , the range of radius included in the fit is indicated by the horizontal line(s ) at the 25.3 @xmath18 level , and the fit parameters are provided near the top of each plot .
this selection of objects includes some very good fits as measured by the rms deviations ( ngc 0016 , ngc 0404 ) , some typical - quality fits ( ngc 0224 , ngc 0237 ) , some fits which avoid strong structure in the profile ( ngc 0151 , ngc 0157 , ngc 0253 ) , and one of the worst fits in the entire sample ( ngc 0289 ) .
there were 39 galaxies in the sample for which no fits were obtained ; these objects are listed in table 2 , with their revised types and axis ratios from the rc3 .
these galaxies are particularly ill - suited for fitting by the selected fitting functions , as many of them have pronounced concavities toward low surface brightness or multiple exponential components in their brightness profiles the chosen functions simply do not represent their light distributions in any meaningful way . finally , there are also several fits ( e.g. , ngc 0016 , ngc 0628 , ngc 0890 , and ngc 5033 ) where the disk component is everywhere fainter than the bulge , and other fits where the disk becomes brighter than the bulge only at intermediate radii ( e.g. , ngc 0670 , ngc 0955 , ngc 1090 , and ngc 1187 ) . in both of these situations ,
the bulge fit might have benefitted from an alternative functional form , perhaps the generalized devaucouleurs law as discussed by andredakis , peletier , and balcells ( 1995 ) , although no attempt has been made to investigate this in the present study .
we have chosen three methods of estimating the fitting errors : 1 ) unweighted rms residuals , 2 ) error estimates provided by the fitting software , and 3 ) comparison with the results of other workers .
the unweighted rms residuals for each fit are tabulated in table 1 . because we have computed them in a uniform fashion for all objects , including all structure in the profiles ,
these values provide an unbiased and consistent measurement of how well the fitting functions and the determined parameters describe the observed brightness distribution . from these values
we find that the median rms deviation of the fits is only 0.15 mag , and that more than 90% of the fits are better than 0.35 mag ; figure 3 shows the distribution of the rms deviations using bins of width 0.05 magnitude .
the overall ability of the fits to describe the brightness distributions is quite good , given that the unweighted rms residuals include all of the structure present in the profile . in an effort to quantify the value of the unweighted residuals as a measure of the goodness of fit
, we have examined plots of the profiles and selected the best examples of profiles without significant structure and which appear to be well represented by the chosen fitting functions . note that this selection did not involve consideration of the computed unweighted rms residuals .
there were 16 objects included in this selection , and they have a mean unweighted rms deviation of 0.05 magnitudes and a range of 0.03 magnitudes to 0.10 magnitudes ; this suggests that fits with unweighted rms residuals greater than about 0.10 magnitudes are affected to some degree by significant structure and/or poor fitting quality .
the full sample contains 161 galaxies with rms residuals of 0.10 magnitudes or less . at the other end of the distribution ,
the worst - fit galaxies have been investigated to attempt to find out why they were so poorly fit .
we have inspected the 18 galaxies which have unweighted rms residuals greater than 0.5 magnitudes to search for common characteristics such as morphology and inclination .
the t - type distribution of these 18 galaxies is essentially the same as for the sample as a whole , so there appears to be no correlation between poor fit and t - type .
there also appears to be no serious trend with inclination : the mean inclination of the group is 57 , consistent with a random distribution of tilts .
the most common characteristic is a low - surface brightness extension to the brightness profile , such that the fit falls below the measurements in the outer portion .
the extension is sometimes featureless , sometimes it contains a distinct bump ( as if an outer ring or arm ) ; sometimes it is nearly constant brightness and other times it is more or less parallel with the inner profile .
there is only one case ( ngc 0157 ) where the problem region is in the main portion of the disk , and this galaxy looks disturbed , almost as if undergoing a collision .
the apparent overriding reason for the poor fits is simply that the chosen fitting functions do not work well for some galaxies .
this same conclusion holds for the 39 galaxies which were not fitted in this effort : common characteristics of these objects are extremely strong , large - scale structures , concavity of the brightness profile toward faint surface brightness , and apparent multiple components of the profile , usually with more than one exponential .
this occasional inappropriateness of the fitting functions suggests that careful consideration of the fitted parameters and their error estimates should be exercised before judging whether a specific fit is truly meaningful for a detailed study of any individual galaxy .
the second method of judging the fitting errors was the use of estimated coefficient errors as produced by the fitting software .
the nfit1d task estimates the errors in each parameter by a process of bootstrap resampling , with a choice of distribution functions for use in the procedure .
we used the poisson distribution as the distribution function for the parameter error estimation as this reflects the photon statistics expected to affect the measured relative intensities in the brightness profiles . in order to minimize the effects of comparing parameters which vary wildly in value from galaxy to galaxy ,
we have computed the fractional error in each fitted parameter .
table 3 lists these fractional error estimates for a sample of the galaxies column 1 gives the galaxy identification and columns 2 through 6 the fractional error of each of the parameters . note that since the fitting was performed in the surface intensity scale ( not surface brightness ) , the relative errors are computed in linear units , not magnitude units . the fractional errors for the full set of galaxies are summarized in table 4 , which gives for each parameter the number of error estimates , the mean fractional error , the standard deviation of the fractional errors , the median fractional error , and the minimum and maximum fractional errors . as can be seen in table 4 , every parameter _ except _
@xmath28 is significantly affected by outliers ; histograms of the fractional error of each parameter are shown in figure 4 , which demonstrate this problem .
removing the most discrepant outlier from the satistical calculations ( second part of table 4 ) improves the results considerably .
unfortunately , there is not just one `` bad '' fit in the sample causing all of these outlying points : the bulge parameter errors are both dominated by the fit for ngc 2441 , while the disk central surface intensity error is affected by the fit for ngc 2541 , the disk scale length error by ngc 5033 , and the hole radius error by ngc 2997 .
the bulge component of ngc 2441 dominates the disk only at radii smaller than about 6 arcsec , so there are very few data points defining the bulge and the coefficient uncertainties reflect this fact . the profile for ngc 2541 is fitted with a faint inner - truncated disk ( the peak disk brightness is only about @xmath37 ) and the hole radius is more than six disk scale lengths from the galaxy center .
the southern half of the brightness profile contains a relatively bright spiral arm ( about @xmath38 at its brightest ) , and this asymmetry results in a bump in the averaged profile which has been fitted with the inner - truncation .
ngc 5033 was fitted with a very faint ( @xmath39 ) , very flat disk ( @xmath29 = 732 arcsec ) which was never comparable in brightness to the bulge . as a result ,
the disk parameters for this galaxy are not well constrained by the data , and the estimated errors in the coefficients are correspondingly large .
finally , ngc 2997 has been fitted with a hole radius just smaller than the radius of the innermost data point point , so the value of the hole radius is , again , not really constrained by the data . because of the presence of these extreme outliers , the median fractional errors are much more useful than the mean for examining the fitting errors of the sample as a whole .
the summary in table 4 shows that the median coefficient fractional errors range from a low of 1.6% for the disk scale length to more than 12% for the bulge effective intensity . from these data , it is clear that the bulge parameters are the least well determined quantities in the fits , while the disk parameters are generally well determined .
this is not unexpected : since the bulge coefficients are usually dominated by a relatively small number of data points , the constraints on them are not very strong .
finally , a few of the galaxies included in this study have been previously fitted by others with the same fitting functions , providing a completely independent check of the results of our fitting procedure .
there are five galaxies in the current study which are also included in boroson ( 1981 ) , and 12 are in common with kent ( 1985 ) , making a total of 17 measurements available for use in this comparison .
prior to making any comparisons between the various works , the surface brightness parameters reported in boroson ( 1981 ) and kent ( 1985 ) have been `` uncorrected '' for the effects of inclination and galactic absorption as applied in each study , and the length parameters have been converted from kiloparsecs to arcseconds using the distances adopted by those authors .
furthermore , boroson ( 1981 ) lists values for the disk @xmath40 color , which have been used to convert his disk central surface brightnesses from the b bandpass to the v. finally , there are two galaxies in common between boroson ( 1981 ) and kent ( 1985 ) , and the same analysis has been applied to them .
the percentage differences between the fitting parameters from these works and the present study have been computed and tabulated in table 5 . in this table
, we list for each galaxy the percentage difference between the parameters in the referenced work and this study , and also between the two reference works ; the last column identifies the reference work .
the differences were computed in the sense reference work minus this study , and boroson minus kent .
these data are also presented in figure 5 , which shows the fitting parameters from boroson ( 1981 ) and kent ( 1985 ) plotted against the values obtained in the present study . in general ,
the fits from the various studies do not agree very well , although the disk fits are typically more similar than the bulge fits .
the mean of the absolute values of the percentage differences are 105% for @xmath26 , 83% for @xmath27 , 61% for @xmath28 , and 26% for @xmath29 , with 17 objects for all parameters .
our bulge fits agree somewhat better with those of kent ( 1985 ) , while our disk central surface brightnesses are closer to boroson s ( 1981 ) results this reflects the conversion of boroson s suface brightnesses to the v - band .
interestingly , our values of @xmath29 agree equally well , on average , with both reference works , leading us to believe that a 25% scatter in the disk scale length is to be expected under the circumstances of this comparison ( differing bandpasses , . the bulge fit differences are dominated by the effects of the profile acquisition procedures used in the different studies : this work used a wedge - shaped major axis cut , while boroson ( 1981 ) and kent ( 1985 ) used two variations on azimuthal averaging . in particular , it has been pointed out by boroson ( 1981 ) that azimuthally averaged profiles will be systematically too bright in the bulge - dominated regions due to sampling the rounder bulge at a smaller galactocentric radius than the disk , at a given position in the image .
inspection of table 5 reveals that every instance of a large departure in @xmath26 shows a large departure in @xmath27 of the _ opposite _ sign . that our bulge parameters agree better with kent ( 1985 ) than with boroson ( 1981 ) probably reflects the use of fixed ellipticities and position angles by boroson , while kent allowed those quantities to vary with radius .
as a result , the mean surface brightness around ellipses in bulge - dominated regions are more representative of the major axis with kent s profile acquisition procedure than with boroson s , hence the slightly better agreement with our major axis cuts .
the basic result of the profile decomposition process is that 620 of the 659 profiles were successfully fitted with the chosen fitting functions . in this section
, we will discuss some of the characteristics of the profile decomposition and the errors as they appear with this sample of galaxies .
our intent is to provide information regarding the characteristics of the fitting ; we will present our analysis regarding the structure of disk galaxies elsewhere .
all galaxies included in this sample have lenticular or spiral rc3 hubble types ; however , 61 of the fits have been made with no disk component , indicating a lack of any appreciable exponential component to their brightness profiles .
visual inspection of these galaxies in the panbg shows many of them to have what appear to be disks even though the profile shows none ( e.g. , ngc 1784 , ngc 1961 , ngc 3370 ) ; others are possibly misclassified elliptical galaxies ( e.g. , ngc 2655 , ngc 3998 , ngc 5485 ) .
the galaxies with no disk component span a range in t - type from -3 to 9 the entire range of t - type included in this sample . early type galaxies are somewhat more likely to be fitted without a disk than are late types , but the trend is not very strong .
there is no significant difference in the fit quality ( as measured by the tabulated rms deviation ) between the galaxies without disk fits and the sample as a whole .
a total of 113 galaxies were fitted without a bulge component , and although these galaxies range in t - type from -2 through 9 , they are mostly late type galaxies . this , of course , is consistent with the basic behavior of the hubble classification scheme where bulges are less prominent in later types .
again , the basic fitting quality is the same for galaxies without bulge fits as with the sample in general .
an investigation of the variation of the median fractional errors with t - type is summarized in table 6 .
this table gives for each range of t - type the median fractional error of each fitting parameter , as well as the number of galaxies included in each median determination .
the errors in @xmath26 , @xmath27 , and @xmath28 show some slight trends which are in the expected senses , but which are small enough to be of questionable significance .
the two bulge parameters seem to have somewhat larger median errors for later types , as would be expected as the bulge contribution to the light distribution decreases .
we should also note the work of andredakis and sanders ( 1994 ) , who show that the inner regions of late - type spirals are perhaps better represented by an exponential light distribution than a devaucouleurs law .
similarly , the disk central intensity error has larger values at earlier types , when the disk contribution is generally lower .
we also looked at the possibility of a fitting quality dependence on the presence or lack of a bar , and we have found nothing significant .
the median rms residual of the fits on non - barred galaxies ( hubble type in the rc3 contains an `` a '' explicitly ) is about 0.14 magnitudes , for barred galaxies ( `` b '' ) it is 0.15 magnitudes , for mixed types ( `` x '' ) 0.16 magnitudes , and for objects with no bar classification given in the rc3 we find a value of 0.14 magnitudes .
all of these values are sensibly the same as as the median value of 0.15 magnitudes found for the full sample , and there is clearly no trend apparent .
table 7 lists the median fractional error of the individual fitting parameters for each bar class ; there are no significant trends in these results .
we conclude that bars have no discernable effect on the quality of the fits .
objects which have been fitted with an inner truncation make up about 25% ( 156/620 ) of the sample . remembering that inner truncations were included only if they improved the _ weighted _ rms by at least 10% , this serves to justify our initial decision to use that function with this large data set .
some of these inner - truncations are probably caused by arms or rings at large galactocentric radii ( e.g. , ngc 2859 , ngc 3368 , and ngc 5701 ) the arm / ring is bright relative to the local disk and thus mimics an inner - truncation . in these cases ,
the arm / ring is typically faint ( peaks near @xmath37 ) , and has a short disk scale length .
these galaxies are also generally classified as having an outer ring .
these objects are identified in table 1 with a `` c '' in the comment column .
a more - detailed analysis of the presence of an inner - truncation is left for a later study .
an analysis of the unweighted rms residuals shows little or no trend in the mean value with inclination .
we divided the sample into three inclination ranges based solely on the @xmath41 value listed in the rc3 , assuming that this isophote corresponds to a flat , circular disk : @xmath42 ( @xmath43 ) , @xmath44 ( @xmath45 ) , and @xmath46 ( @xmath47 ) .
note that no t - type dependence was included in this inclination estimate .
the resulting mean rms residuals for the low inclination , medium inclination , and high inclination samples are 0.16 , 0.19 , and 0.21 magnitudes , respectively .
these three values are all much less than one standard deviation from each other , so the trend is statistically meaningless .
the median rms residuals for each inclination sample are 0.15 magnitudes , 0.14 magnitudes , and 0.17 magnitudes for the low , medium , and high inclinations , respectively .
we conclude that the sample has no significant inclination dependence on the unweighted rms residuals of the fits .
the quality of the individual fitting components is investigated by computing the median fractional errors within each inclination range ; the mean fractional errors are not useful because of the outliers discussed previously .
table 8 provides the results of this investigation , which shows that there are no indications of any inclination dependencies . in table 8
, the first column lists the parameters , and the next three columns give the median relative errors in each inclination range . finally , we check if the rate of occurrence of the inner - truncation has any inclination dependence by computing the rate in each inclination range .
the rates are @xmath48 for the low inclination group , @xmath49 for the medium inclinations , and @xmath50 for the high inclination sample .
thus , there is no trend significant at the @xmath51 level .
we have presented one of the largest , if not the largest , collections of spiral and lenticular galaxy brightness profile bulge - disk decompositions yet completed . of the 659 brightness profiles in our sample ,
620 were fit with the devaucouleurs law plus inner - truncated exponential disk function , while the remaining 39 profiles could not be so fit .
the general quality of the fits is quite high , with about 50% of the fits having an unweighted rms deviation from the data ( including real structures ) of less than 0.15 magnitudes and more than 90% have unweighted
rms residuals of less than 0.35 magnitudes .
we find no systematic trends in the fitting quality with either galaxy morphology or inclination .
comparison of our fits with those of boroson ( 1981 ) and kent ( 1985 ) show discrepancies attributable to a number of observational and data reduction factors .
probably the most interesting result from this process of fitting is simply that we achieved a `` success '' rate of 620/659 ( 94% ) in our fits , compared to success rates of 75/94 ( 80% ) for kent ( 1985 ; spiral and s0 galaxies only ) and 16/26 ( 62% ) for boroson ( 1981 ) . while these success rates are statistically similar , we wonder if the small number statistics are the only differences .
an analysis of the 19 non - fittable disk galaxy profiles from kent ( 1985 ) shows that we found fits for all ten of those galaxies which were also in our sample . for these 10 objects ,
the mean inclination is 41@xmath52 18@xmath53 , the median is 42@xmath53 , and only one galaxy ( ngc 5566 ) has an inclination greater than 60@xmath53 .
thus , kent s non - fittable galaxies do not tend to be high inclination objects .
furthermore , the rms deviations of our fits for these same galaxies are generally small , with a median value of 0.15 magnitudes , the same as for our full sample . the princpal difference seems to be that we included the inner - truncated disk ( it d ) factor in our fits : seven of these 10 galaxies have it d s in our fits , often with large @xmath54 ratios .
it is also possible that kent s simultaneous fitting of the minor axis profiles made his results more sensitive to deviations from the standard fitting functions . a similar analysis of the 10 non - fittable profiles in boroson s ( 1981 ) sample shows us having fits for the nine which are included in our sample .
these nine galaxies are also of relatively low inclination ( the largest is about 58@xmath53 ) , and have small rms deviations in our fits ( median value of 0.13 magnitudes ) . however ,
our fitted parameters show three objects with bulge only , three with a bulge plus exponential disk , and three with a bulge plus it d ; the case for the inclusion of the it d is not as strong with this set of profiles .
we conclude , however , that the inclusion of the it d term in our fitting functions has allowed us to fit an additional 10% to 15% more galaxies than we would have fit without the inner - truncation term .
it is also interesting that about 25% of the profiles in our sample are fitted better by an inner - truncated disk function than with a plain exponential .
some of these fits are certainly due to the it d being fitted to outer rings ( as indicated by a `` c '' in the last column of table 1 ) , and others may be marginal improvements ( remember the requirement for a 10% improvement of the _ weighted _ rms to include the inner - truncation term ) , but clearly a significant fraction of the profiles support the physical reality of the inner - truncation in the light distribution .
a quantification of the strength of the inner truncation and the search for the origin of this feature is the subject of a future paper .
acknowledgements : one of the authors ( web ) has been supported by stsci under contract nas5 - 26555 for this work .
the authors would like to thank m. hamabe for making the panbg brightness profiles available to us for this project .
we also wish to thank the anonymous referee for some useful suggestions .
part of the data analysis for this paper used stsdas , which was developed by the space telescope science institute under u.s .
government contract nas5 - 26555 .
the fitting procedure described in section 3.1 makes no allowance for the effects of seeing other than to start the fitting at a radius larger than @xmath55 to avoid the most - affected portion of the brightness profile .
we describe in this appendix a set of experiments which were used to derive estimates of the correction factors for the fitted parameters to measure more accurately the true parameters of the bulge light distributions .
there are four factors in the profile acquisition process which can influence the fitting results , all of which occur at distinct stages of the process and which can be assumed to be separable . seeing smears the galaxy light as it travels through the atmosphere , while poisson noise occurs during the photographic exposure physics and chemistry .
pixellation broadens sharp features during the digitization of the plate , as well as adding some additional poisson noise , and smearing by the aperture photometry happens as a result of the profile acquisition from the digitized data .
our experiments were designed to investigate only those processes which broaden sharp features of the galaxy light distribution : seeing , pixellation , and aperture photometry .
the panbg provides seeing estimates for all plates in their table 4.1 ; we assume that `` seeing '' in this case refers to the fwhm of stellar profiles .
the tabulated seeing in the panbg ranges from @xmath56 to @xmath57 and is included in our table 1 .
we attempted to replicate the profile definition process as accurately as possible utilizing computer - generated images .
these images consist of a perfectly circular devaucouleur s light distribution with input values of @xmath58 in arbitrary intensity units and @xmath59 and @xmath60 .
poisson noise was included in the images .
no disk component was included in these images because the most pronounced effects of seeing will be on the bulge component due to its very steep slope at small radii .
we generated 512x512 pixel images with these characteristics , assuming a pixel scale of 3 pixels/@xmath61 ( to simulate the photographic resolution ) , then smeared them with gaussians of fwhm=1 , 2 , 3 , 4 , 5 , 6 , and @xmath57 to simulate the effect of the atmosphere during a long exposure .
the resulting images were then block averaged with a 3x3 pixel ( @xmath56 square ) block to replicate the plate scanning utilized for most of the galaxies in the panbg .
finally , the brightness profiles from the simulated images were extracted using a set of variable diameter circular apertures along a single radius of the `` galaxies . ''
these profiles were then fitted in the same fashion as the real profiles in order to estimate the functional parameters , using a fixed fitting range of @xmath55 to @xmath62 .
a second set of images was generated using the same input parameters , but which were not smeared by the gaussians ; these images were used to determine the correction factors due to the pixellation and profile acquisition aperture photometry prior to estimating the effects of smearing with a gaussian . the results of this procedure are presented in table 9 , which includes for each simulation the seeing size in arcseconds , the input @xmath35 and @xmath36 , the fitted @xmath35 and @xmath36 , and the ratio of each fitted parameter to its input value , all based on the fully - degraded brightness profiles .
the @xmath35 are all in arbitrary intensity units , and the @xmath36 are all in simulated arcseconds .
figure a1 shows the relationship between the fitted : input @xmath35 as a function of the fitted : input @xmath36 ; the two quantities are highly anti - correlated , suggesting that the fitted parameters are not truly independent of one another . as can be seen from the table ,
there are some cases where the fitted parameters are very different from the input values : the worst cases , as would be expected , are where the seeing disk is large compared to the input @xmath36 . for @xmath36 ,
the percent errors range from less than 1% to almost 40% ( seeing=@xmath57 , @xmath63 ) , while for @xmath35 the range is from almost 0% ( seeing=@xmath64 , @xmath65 ) to more than 1000% ( seeing=@xmath57 , @xmath63 ) .
we have every reason to suspect that similar fitting errors exist in the panbg fits , and a means of correcting these errors is highly desirable . the correction procedure has been separated into two stages : first correct the fitted parameters for the combined effects of pixellation and aperture photometry , then correct these modified parameters for the effects of the gaussian smearing .
the pixellation / aperture photometry correction for the effective radius , based on the simulations without gaussian smearing , has been found to be well - represented by a power law of the form where @xmath4 is the fitted effective radius , @xmath68 is the digitization aperture size ( @xmath56 ) , and @xmath69 is the effective radius corrected for the effects of pixellation and aperture photometry ; the units of all quantities are in arcsec .
this function has a correlation coefficient of 0.989 , and a maximum percentage error of 0.2% within the parameter space studied . where @xmath5 is the fitted effective intensity
, @xmath4 is the fitted effective radius , @xmath68 is the digitization aperture size , and @xmath71 is the effective intensity corrected for pixellation and aperture photometry .
this function has a correlation coefficient of 0.995 and a maximum percentage error of 0.6% within the parameter space studied .
figure a2 shows the fits of the pixellation / aperture photometry correction functions for both @xmath4 and @xmath5 .
the seeing correction is applied by the linear interpolation of the galaxy parameters on the model parameters for the same value of seeing . in table 10
we list the model data used in the interpolation ; column 1 lists the seeing in arcsec ( @xmath72 ) , column 2 the ratio of the seeing to the pixellation - corrected effective radius ( @xmath73 ) , column 3 the ratio of the input effective intensity to the pixellation - corrected effective intensity ( @xmath74 ) , and column 4 the ratio of the input effective radius to the pixellation - corrected effective radius ( @xmath75 ) .
the interpolation procedure is performed by computing the @xmath73 ratio for a fit , and interpolating the @xmath76 and @xmath77 ratios for the tabulated seeing disk size to derive the fully - corrected @xmath78 and @xmath79 from the previously - computed values of @xmath71 and @xmath69 ( eqs .
a2 and a3 ) .
we include in table 1 the corrected bulge parameters resulting from the described correction procedure .
corrections for objects whose fitted parameters lie outside of the interpolation parameter space are not provided in table 1 .
andredakis , y.c . ,
peletier , r.f . , and balcells , m. 1995 , , 275 , 874 andreyakis , y.c . and sanders , r.h .
1994 , , 267 , 283 baggett , w.e .
, baggett , s.m . , and anderson , k.s.j .
1993 , , 24 , 1223 bevington , p.r .
1969 , _ data reduction and error analysis for the physical sciences , _
( mcgraw - hill book company , new york ) , p 184 boroson , t. 1981 , , 46 , 177 burstein , d. 1979 , , 234 , 435 byun , y.i . , and freeman , k.c .
1995 , , 448 , 563 dejong , r. 1996 , , 313 , 45 freeman , k.c .
1970 , , 160 , 811 holmberg , e. 1947 , medd .
2 , no . 120 hubble , e. 1930 , , 71 , 231 kent , s. 1985 , , 59 , 115 king , i.r .
1966 , , 71 , 64 kodaira , k. , okamura , s. , and ichikawa , s. ( eds . ) 1990 , _ photometric atlas of northern bright galaxies , _ ( university of tokyo press , tokyo ) ( panbg ) kormendy , j. 1977 , , 217 406 longo , g. and devaucouleurs , a. 1983 , _ the university of texas monograph in astronomy , _ no .
3 , ( university of texas , austin ) sandage , a. 1961 , _ the hubble atlas of galaxies , _ ( carnegie institution of washington , washington ) sandage , a. and tammann , a. 1981 , _ a revised shapley - ames catalog of bright galaxies , _ ( carnegie institution of washington , washington ) ( rsa ) devaucouleurs , g. 1953 , , 113 , 134 devaucouleurs , g. , devaucouleurs , a. , corwin , h.g . ,
buta , r.j . ,
paturel , g. , and fouque , p. 1991 , _ third reference catalogue of bright galaxies , _ ( springer - verlag , new york ) | we present one of the largest homogeneous sets of spiral and lenticular galaxy brightness profile decompositions completed to date .
the 659 galaxies in our sample have been fitted with a devaucouleurs law for the bulge component and an inner - truncated exponential for the disk component .
of the 659 galaxies in the sample , 620 were successfully fit with the chosen fitting functions .
the fits are generally well - defined , with more than 90% having rms deviations from the observed profile of less than 0.35 magnitudes .
we find no correlations of fitting quality , as measured by these rms residuals , with either morphological type or inclination .
similarly , the estimated errors of the fitted coefficients show no significant trends with type or inclination .
these decompositions form a useful basis for the study of the light distributions of spiral and lenticular galaxies .
the object base is sufficiently large that well - defined samples of galaxies can be selected from it . |
driven interfaces in random media display intriguing scaling laws that are common to a wide variety of phenomena , including fluid imbibition , crack front roughening , dislocation hardening , superconducting flux lines , the equilibrium motion of piles of rice down an incline , and domain wall motion in magnets @xcite .
the scaling laws are commonly associated with an underlying depinning critical point that has been elucidated by simple models for interface dynamics .
these models have been extensively studied using continuum simulations @xcite , cellular automata @xcite , and field - theoretic @xmath0 expansions @xcite , providing a sophisticated picture of the non - equilibrium phase transition and of the different universality classes .
the interface morphology is usually characterized by the roughness exponent @xmath1 , resulting from a coarse graining operation of the interface height function @xmath2 .
namely , when we change all length scales by a factor @xmath3 , or @xmath4 , then statistically @xmath5 hence @xmath6 for many experiments and simulations , it is convenient to measure @xmath1 by computing the height - height correlation of the interface @xmath7 ( in sec . ii @xcite we shall study a system with _ anomalous scaling _ , where the power law exhibited by @xmath8 differs from the universal rescaling exponent @xmath1 . rather than rescaling @xmath9 in such systems ,
one studies the rescaling of the correlation function directly , @xmath10 . these systems are multi - affine @xcite : different moments of @xmath9 will scale with different exponents . ) here @xmath1 should be uniquely determined by which universality class the system belongs to .
however , in practice , the observed @xmath1 varies ( see table [ table : roughness ] ) even for the same type of system , such as paper wetting .
measuring a single exponent for these systems may prove inadequate due to the presence of _ crossover behavior _ between universality classes .
this is a common source of confusion and controversy .
if the crossover is gradual , an experiment or simulation may measure an effective exponent @xmath11 intermediate between existing theories , and appear to demand a new theoretical explanation ( _ i.e. _ , universality class ) .
c c c experiment & @xmath1 & reference fluid flow & 0.73 & @xcite & 0.81 & @xcite & 0.65 - 0.91 & @xcitepaper wetting & 0.63 & @xcite&0.62 - 0.78 & @xcitebacteria growth & 0.78 & @xciteburning fronts & 0.71 & @xcite in section [ sec : fracture ] we analyze a straightforward experimental example of a crossover between two forms of roughness in two - dimensional fracture .
there we introduce the universal crossover scaling functions , and provide a brief renormalization - group rationale . in the remainder of the paper
, we examine a more complex theoretical model .
crossovers , long studied in ordinary critical phenomena , have now been studied for several interface models @xcite , however theoretical studies have proven challenging in different ways @xcite . for thin film magnets , the experiments
@xcite observe a crossover between short - range and mean - field universality classes as long - range dipolar fields are introduced , which can be done by changing the thickness of the film .
however , for models of that type , simulations are challenging , both because of the long - range fields and the striking zig - zag morphologies that emerge and compete with the avalanche behavior .
crossovers involving the transition between depinning and sliding dynamics incorporating periodically correlated disorder @xcite have also been studied .
it is not typical , however , to study and report the universal scaling functions for these crossovers a challenge we now shall address .
we shall analyze a numerically tractable , but analytically tricky crossover @xcite : the transition between the linear , super - rough , quenched edwards - wilkinson model ( qew ) and the nonlinear quenched kpz model ( qkpz ) @xcite . in both experiment and theory ,
we focus on the crossover behavior of the height - height correlation function .
of a 2d fracture front , generated by pulling apart two pieces of pmma that have been sand - blasted and sintered together @xcite .
the three curves differ in the size of the sand - grain beads ; the relation between the bead size and the toughness fluctuations in the pmma were not measured .
the dashed lines show two different power - law critical regimes , with @xmath12 and @xmath13 , governing the short- and long - distance scaling behavior : the crossover between these regimes is evident . here
our fit gives @xmath14 and @xmath15 , within the experimentalists suggested range @xmath16 and @xmath17 .
the main figure shows a scaling plot of @xmath18 versus @xmath19 , with the curves shifted vertically and horizontally to best collapse .
the thick black curve is a one - parameter fit of the universal scaling function to the functional form in eqn ( [ eq : fracturecscaling ] ) . ] just as the critical exponent @xmath1 is universal ( independent of microscopic details , within a class of physical system ) , so too is the crossover behavior between universality classes . as a simple example , santucci et al .
@xcite have measured a relatively sharp crossover between two regimes for two - dimensional fracture ( inset in fig .
[ fig : fracture ] ) . because the fracture is done slowly
, we can view the crack front as self - organizing to the depinning transition for the crack front .
well below a critical distance @xmath20 , they observe a power law @xmath21 with an exponent that was interpreted as originating from coalescing cracks @xcite or with larkin scaling @xcite .
well above @xmath20 they observe a different power law @xmath22 consistent with the depinning transition of a line @xcite .
the crossover between these two universal power - law regimes should be described by a _
universal crossover function _
@xcite , @xmath23 : @xmath24 independent of microscopic details . at small arguments
@xmath25 must go to a constant , and at large arguments @xmath26 , so as to interpolate between the two power laws . when analyzing different systems governed by the same universal crossover , one may plot all the crossovers in a scaling plot , dividing the distances @xmath19 on the ordinate by a system - dependent factor @xmath20 for each curve , and dividing the magnitudes of the correlations on the abscissa by a system dependent constant @xmath27 ( see fig . [
fig : fracture ] ) .
the resulting data curves then should align , giving the universal function @xmath28 . to continue with this simple test case
, we may fit the universal scaling function to an approximate functional form .
( indeed , we find it convenient to do a joint fit of the functional form , the exponents , and the constants @xmath20 and @xmath27 . ) to the extent that a guessed functional form reproduces the universal one , it is equivalent : advanced field - theoretic methods for calculating exact scaling functions are nt needed to analyze future experiments . however , judicious choices of functional forms with the correct limiting behavior can greatly facilitate this process .
the interpolation @xmath29 has the correct limits , but its rather gradual crossover does not explain the data . we may heuristically add a parameter @xmath30 which at large values produces an abrupt crossover : @xmath31 this yields an excellent fit to the data with @xmath32 ( see fig .
[ fig : fracture ] ) .
why is the scaling form of eqn ( [ eq : cfracscaling ] ) expected ?
briefly , the renormalization group studies the behavior of systems under coarse - graining : describing the properties of a system at length scales changed by a factor @xmath3 .
one gets universal power laws when the system becomes invariant under repeated coarse - grainings : if @xmath33 , under coarse - graining by a factor @xmath3 , then by coarse - graining @xmath30 times such that @xmath34 one has @xmath35 . in the case of a crossover ,
a fixed point is unstable to some direction @xmath36 in system space .
then a small initial @xmath36 grows under rescaling by some factor @xmath37 , so @xmath38 .
now rescaling until @xmath39 , we have @xmath40 where we choose @xmath41 .
if the unstable direction flows to a new fixed point with a different @xmath42 , that behavior will be reflected in the large - x dependence @xmath43 ( * ? ? ?
* section 4.2 ) .
note that different physical systems will have different overall scales of height fluctuations , so we must have an overall scale @xmath27 for each experiment .
( if the experiments fall into a parameterized family , @xmath27 will depend smoothly on the parameters , giving _
analytic corrections to scaling _ as discussed in section [ sec : results ] . )
note , though , that the rescaling factor @xmath20 for lengths , while it will still vary from one system to another , now depends on @xmath36 with a power - law singularity , as @xmath44 ; within the renormalization group , @xmath36 measures how far along the unstable direction the original system was poised . in particular
, @xmath20 becomes large as @xmath45 , as in that limit the unstable fixed point remains in control .
the three experiments depicted in fig . [ fig : fracture ] started with different bead sizes .
if all other features of the experiment are held fixed , one may assume that the control parameter @xmath36 depends in some smooth way on the bead size .
had we several values of bead size , we could then extract values for the universal crossover exponent @xmath46 . in the following sections
, we shall perform a far more sophisticated version of this type of analysis . by exhaustively varying system size and nonlinearity in an interface growth model
, we shall not only generate universal _ two - variable _ functional forms for the correlation crossover scaling function , but will be able to make predictions about both the dependence of the crossover length scale ( corresponding to @xmath20 ) and the dependence of the correlation amplitude ( corresponding to @xmath27 ) on the control parameters . a rich , nuanced understanding of the model behavior thus emerges .
the equations of an interface in a disordered environment may be written generally as follows .
let the one - dimensional interface , @xmath47 be driven by a force @xmath48 through a disordered environment with a local quenched random force @xmath49 : @xmath50 here @xmath51 is a surface tension , and @xmath36 is the coefficient of the kpz term .
the kpz term controls lateral spreading of the interface , breaks the statistical tilt symmetry , and changes the universality class @xcite .
@xmath48 is a slowly - increasing external driving force .
our simulations are done with a lattice automaton ; the lattice naively might be thought to break this statistical tilt symmetry , but simulations have long shown that the model faithfully describes both universality classes @xcite .
the term @xmath52 is borrowed from simulations of magnets , where it represents the demagnetization force @xcite , approximating the effects of the long - range dipolar field cost of a net advance in the front .
this restoring force ` self - organizes ' the depinning transition to the fixed point , allowing simulations to access many metastable states , without having to enforce an actual quasi - static field .
it is known @xcite that this restoring force does not produce loop corrections to the renormalization group equations and therefore does not change the universality class of the problem .
we have confirmed numerically that its effects are small for our crossover and appear irrelevant . as the restoring force makes the simulation vastly more efficient
, we include this restoring force , but we do not include @xmath53 in our scaling analysis .
using the automaton simulation employed in @xcite , we tune @xmath54 from 0 to 5 , and observe how the resulting behavior changes . figure [ fig : crossover ]
shows how the front morphology qualitatively changes while we increase the nonlinear parameter @xmath36 .
notice that with increasing @xmath36 the fronts between events are flatter than at small @xmath36 .
256 simulations with the nonlinear kpz term coefficients set to ( a ) @xmath55 , ( b ) @xmath56 , ( c ) @xmath57 , ( d ) @xmath58 .
the random colors / shades represent the area between each pinned front .
one can see that the morphology of the interfaces change dramatically as @xmath36 increases . ] according to equation [ eq : heightzeta ] , naively one would assume we could recover the exponent @xmath1 by defining an effective exponent @xmath11 to be half the local - log slope of the height - height correlation functions ( figure [ fig : local_log_new ] ) . from other numerical studies , for qew , we expect @xmath59 ( cellular automata @xcite models show @xmath60 ; continuous string models @xcite found @xmath61 . ) . for qkpz
, we expect @xmath62 @xcite .
however , there are two things about figure [ fig : local_log_new ] worth noting : ( 1 ) the slope - measure of @xmath1 drifts between @xmath63 and @xmath64 as we change @xmath36 , ( 2 ) the measured value is never greater than one as is naively expected for the linear qew model .
the dropoff at @xmath65 is due to the periodic boundary conditions . of the height - height correlation function for varying @xmath36 and @xmath53 .
the lower dashed red line is @xmath62 as expected for the kpz universality class .
the upper dashed blue line is @xmath66 , the largest growth allowed for super - rough interfaces .
the thin lines show the predictions of our fits ( section [ sec : results ] ) , and the dashed black line is the fit value of @xmath67 .
the sharp cutoff in the curves at large @xmath19 is due to the periodic boundary condition , that forces @xmath68 at @xmath69 ; the larger @xmath70 curves have later cutoffs .
note as expected that for small @xmath36 the curves are described by the ew behavior @xmath71 at small @xmath19 , while for large @xmath36 they are well described by @xmath67 . for intermediate values of @xmath72
, we observe a clear transition from ew behavior at small @xmath19 to @xmath73 behavior at intermediate @xmath19 , before being cut off by the finite size effects . ]
-0.9truein the second issue has a known resolution : for @xmath74 , when the interface is ` superrough ' , the height can not grow faster than linearly with distance , so the height - height correlation function can not directly exhibit a power law larger than one @xcite .
this so - called anomalous scaling @xcite implies that the exponent @xmath1 is reflected not in the distance - dependence of the correlation function , but rather in its _ system - size dependence_. we thus consider the finite - size scaling form @xmath75 the roughness exponent @xmath76 may be estimated by the system - size dependence of the magnitude of @xmath77 .
note that the periodic boundary conditions implies that @xmath78 ; near @xmath79 the correlation function reaches a peak ( and @xmath80 vanishes , as in fig .
[ fig : local_log_new ] ) .
thus @xmath81 . to control the sharpness of the peak in the correlation function at @xmath82 , in analogy to the crossover sharpness parameter @xmath30 of eqn [ eq : fracturecscaling ]
, we introduce @xmath83 giving a transition between the two power - laws : @xmath84 for qkpz ( fig .
[ fig : crossover]c ) , the correlation function in a system size @xmath70 takes the finite - size scaling form @xmath85 where we introduce @xmath86 to form a periodic functional form @xmath87 ( black curves ) with fit parameters @xmath88 , @xmath89 , @xmath90 , @xmath91 , @xmath92 .
@xmath93 , @xmath94 , @xmath95 , @xmath96 , and @xmath97 .
( a fit constrained to the expected values of @xmath76 and @xmath67 give slightly worse , but acceptable fits , with the other parameter estimates within the quoted ranges . )
the legends denote @xmath70 and @xmath36 for each simulated correlation function ; all runs had @xmath98 .
the errors quoted are a rough measure of the systematic error @xcite , as described in the text , and are representative of the differences we find using different weightings and functional forms .
( they are much larger than the statistical errors . )
note that the amplitude dependence is captured by the scaling form .
three of the twelve parameters ( @xmath76 , @xmath67 , and @xmath46 ) are universal critical exponents , three ( @xmath99 , @xmath100 , and @xmath101 ) describe the non - universal dependence of an overall height scale on parameters , two describe finite - size effects , and only two ( @xmath102 and @xmath103 ) are needed to describe the universal crossover function to the accuracy shown . ]
-0.5truein \(a ) to illustrate the crossover between ew and kpz - dominated lengths ( @xmath104 and @xmath105 respectively , see also @xcite ) .
here @xmath106^{-1/(2 - 2{{\zeta_{kpz}}})}$ ] is the distance where the ew and kpz components of the correlation function are equal in magnitude , and @xmath107 factors out the dependence expected in the kpz regime ( hence yielding flat behavior for @xmath108 ) . in ( b ) , we also factor out the effects of analytic corrections to scaling ; here all of the curves lie on a universal curve apart from the effects of the finite system sizes ( causing each curve to drop on the right ) . in ( a )
, we see that the analytic correction to scaling has a significant impact on the scaling collapse : ignoring it in the analysis would produce significant errors in critical exponents and scaling functions .
the thick blue curve shows the scaling function prediction for the crossover ( which , if @xmath109 , can be shown to be @xmath110 ) .
, title="fig : " ] ( b ) to illustrate the crossover between ew and kpz - dominated lengths ( @xmath104 and @xmath105 respectively , see also @xcite ) . here
@xmath106^{-1/(2 - 2{{\zeta_{kpz}}})}$ ] is the distance where the ew and kpz components of the correlation function are equal in magnitude , and @xmath107 factors out the dependence expected in the kpz regime ( hence yielding flat behavior for @xmath108 ) . in ( b ) , we also factor out the effects of analytic corrections to scaling ; here all of the curves lie on a universal curve apart from the effects of the finite system sizes ( causing each curve to drop on the right ) . in ( a )
, we see that the analytic correction to scaling has a significant impact on the scaling collapse : ignoring it in the analysis would produce significant errors in critical exponents and scaling functions .
the thick blue curve shows the scaling function prediction for the crossover ( which , if @xmath109 , can be shown to be @xmath110 ) .
, title="fig : " ] -0.3truein the drift in the exponent @xmath1 , however , demands a study of the scaling near the unstable qew fixed point , and the functional form of the resulting crossover scaling function .
the role of @xmath36 in generating the crossover from qew to qkpz has only been studied qualitatively @xcite , with no full description of the crossover scaling @xcite .
the crossover describes the rg flow from the qew fixed point to the qkpz as the relevant parameter @xmath36 is added .
the scaling form for the height - height correlation function is thus that of a relevant variable @xmath36 added to the qew scaling : @xmath111 for @xmath112 , @xmath113 , therefore , @xmath114 here @xmath115 is in general a non - universal prefactor for the @xmath73 correlation function . as with @xmath27 in section [ sec : fracture ] , @xmath116 is expected to vary is presumably not the natural measure of the unstable direction in parameter space ; in general that would be an analytic function @xmath117 of @xmath36 and other parameters . ] with the parameters of the problem .
here it has a typical smooth variation @xmath118 , times a singular piece @xmath119 due to the ew fixed point : @xmath120 .
we derive the power - law divergence of the amplitude @xmath121 by noting that @xmath122 must be a scaling function with only invariant combinations of @xmath19 , @xmath70 , @xmath36 .
we must also have @xmath123 .
using these limits , equation [ eq : crosscurlyc ] , and the fact that @xmath115 gets large as @xmath124 and small as @xmath125 , we can construct a function that crosses over between these two limits : @xmath126 where @xmath127 and @xmath128 , and @xmath129 and @xmath77 are defined in eqns .
( [ eq : cew]-[eq : ckpzform ] ) .
we expand the analytic function @xmath130 in a form analytic at zero and saturating at large @xmath36 at @xmath101 , and @xmath131 indicate a saturation of the amplitude of the kpz power - law scaling .
this suggests that @xmath132 at larger @xmath36 . in the range we consider
, an asymptotically flat asymptote suffices to capture this variation ; independent fits of the amplitude at each @xmath36 including the larger systems give comparable fits and other parameters within our ranges . ] and we include a relative scale factor @xmath103 .
finally , we vary the sharpness of the crossover with @xmath102 , just as we did in the experimental study of fracture ( eqn [ eq : fracturecscaling ] ) .
the theoretical curves in were fit to the data in figs .
( [ fig : local_log_new ] ) and ( [ fig : corrfit ] ) and ( [ fig : sofq ] ) , deleting the noisy half near @xmath69 in the first , and using weights @xmath133 designed to equalize the emphasis on each decade .
the errors quoted are a rough estimate of the systematic error @xcite given by quadratically exploring fits with roughly twice the @xmath134 of the best fit .
note that this gives us a universal function of three variables ( @xmath19 , @xmath70 , and @xmath36 ) .
note that it predicts a singularity at small @xmath36 in the form of a divergent amplitude @xmath135 in the qkpz correlation function ( eqn [ eq : ckpz ] ) as @xmath136 .
this universal singularity in the amplitude ( corresponding to a prediction of @xmath27 in section [ sec : fracture ] ) explains the amplitude dependence seen in fig .
[ fig : corrfit ] .
there is an analogous universal amplitude dependence seen for the heisenberg@xmath137ising crossover at small ising anisotropy ( * ? ? ?
* section 4.1 ) . for the height fluctuations .
note that rescaling by the naive rg power @xmath138 here does collapse the data for vanishing @xmath36 ( up to a significant noise ) : the anomalous scaling for the super - rough interface in real space is not manifested in fourier space .
the theory curves are given by the fourier transform of the same best fit shown in previous figures . ]
-0.5truein we can use the scaling form of the correlation function @xmath139 to derive other , less complex crossover scaling functions traditionally studied in interface depinning problems @xcite . the spectral scaling function is equal to the fourier transform of our correlation function ( with subtleties at @xmath140 ) : @xmath141 here the universal spectral crossover scaling function @xmath142 can be written in terms of our universal correlation crossover scaling function ( @xmath143 from eqn [ eq : ccross ] ) as @xmath144 depicted in fig .
( [ fig : sofq ] ) . here
it is known that @xmath145 does not have anomalous scaling ; the power laws @xmath76 and @xmath67 can be read off from the slopes at large and small @xmath146 .
more ordinary , single - variable scaling functions can be derived from our two - variable scaling functions , such as that governing the average width of an interface ( given by the zero fourier component of @xmath147 or an integral of @xmath8 ; such scaling functions ( as for the fracture scaling function of eqn [ eq : cfracscaling ] ) allow traditional scaling collapses ( as in fig .
[ fig : fracture ] ) .
however , one should note that our analyzed simulations extend to @xmath148 , where analytic corrections to scaling , as vividly illustrated in fig .
[ fig : corrcollapse](a ) , would likely invisibly distort the resulting scaling collapses .
it is an advantage of multivariable scaling fits that they both allow the incorporation of such analytic corrections ( extending the range of applicability ) , and force their incorporation ( exposing weaknesses of the naive theory ) .
in this paper , we have analyzed the scaling properties for an experimental 2d fracture front and a model of an interface moving in random media , focusing on the crossover scaling of the roughness .
the experimental system is successfully modeled using a one - variable universal scaling function with one free parameter , controlling the sharpness of the transition .
the theoretical model , the crossover from the qew to the qkpz universality class with the addition of a non - linear term , allows us to estimate the complete universal scaling function for the height - height correlation function including both finite - size effects and the non - linear effects of the tuning parameter @xmath36 , while satisfying known limits given by the renormalization group .
we emphasize here the importance of our sophisticated use of the scaling forms and corrections predicted from the renormalization group .
[ fig : fracture ] illustrates that not only the power - laws , but the entire functional form of the crossover , is a universal property that should be reported . equation [ eq : fracturecscaling ] is an effective one - parameter way of embodying the sharpness of the crossover , which we use also in the theoretical analysis of section [ sec : results ] for both the crossover and the effects of periodic boundary conditions .
figures [ fig : local_log_new ] , [ fig : corrfit ] , and [ fig : sofq ] show how different experimental characterizations of the roughness of an interface can be simultaneously fit with a single functional form .
figure [ fig : corrcollapse](a ) vividly indicates the importance of analytic corrections to scaling in extending the validity of the theoretical analysis to smaller systems and farther outside the critical region .
only systems with @xmath149 will follow the scaling behavior without incorporating the analytic corrections , while the entire crossover is faithfully represented in fig .
[ fig : corrcollapse](b ) by including them in the fit . by developing functional forms for the correlation functions @xcite
, we gain the flexibility of incorporating analytic corrections , multiple scaling variables , and a systematic error analysis while allowing the quantitative reporting of the universal scaling functions .
one should note that the parameter estimation errors quoted here are large compared to more traditional scaling analyses . in part this is due to our estimation of the relevant systematic errors @xcite ; statistical errors would be perhaps an order of magnitude smaller . in part , however , this is due to our incorporation of known but usually ignored confounding factors analytic corrections to scaling and crossover effects will invisibly distort the results of more direct measurements , and the drift in exponents quoted in the literature in critical phenomena often exceeds the error estimates .
it is challenging but satisfying to develop these functional forms . measuring and fitting them
is far more physically intuitive and less technically demanding than direct field - theoretic calculations @xcite ( although theoretical calculations often form important inspiration for choosing functional forms @xcite ) .
successful functions are parsimonious in the number of adjustable parameters , and developing them often forces one to develop a far more complete understanding of the physics of the system under consideration .
this work is supported by nsf and cnr through materials world network : cooperative activity in materials research between us investigators and their counterparts abroad in italy ( nsf dmr 1312160 ) , and through nsf phy11 - 25914 .
sz acknowledges support from the academy of finland fidipro program , project 13282993 and erc advanced grant sizeffects .. | we study the crossover scaling behavior of the height - height correlation function in interface depinning in random media .
we analyze experimental data from a fracture experiment and simulate an elastic line model with non - linear couplings and disorder . both exhibit a crossover between two different universality classes . for the experiment
, we fit a functional form to the universal crossover scaling function . for the model ,
we vary the system size and the strength of the non - linear term , and describe the crossover between the two universality classes with a multiparameter scaling function .
our method provides a general strategy to extract scaling properties in depinning systems exhibiting crossover phenomena . |
many organizations rely on the skills of innovative individuals to create value .
examples include academic institutions , government organizations , think tanks , and knowledge - based firms .
workers in these organizations apply a variety of skills to in order to solve difficult problems : architects design buildings , biochemists develop new drugs , aeronautical engineers create bigger and better rockets , software developers create new applications , and industrial designers create better packaging materials .
their success and thus the success of the organizations they work for is dependent on the particular set of skills that they have at their disposal , but in most cases , the decision of which skills to acquire is made by individuals , rather than organizations .
the perception is that these workers choose to become more specialized as the problems they face become more complex ( strober ( 2006 ) ) .
this perception has generated a countervailing tide of money and institutional attention focused on promoting interdisciplinary efforts .
however , we have very little real understanding of what drives an individual s decision to specialize .
roughly speaking , workers in knowledge - based fields can be divided into two categories : specialists , who have a deep knowledge of a single area , and generalists , who have knowledge in a wide variety of areas . in this paper ,
i consider an individual s decision to be a specialist or a generalist , looking specifically at two previously unaddressed questions .
first , under what conditions does it make sense for an individual to acquire skills in multiple areas ? and second , are the decisions made by individuals optimal from an organizational perspective ?
most of the work done on specialists and generalists is focused on the roles the two play in the economy .
collins ( 2001 ) suggests that specialists are more likely to found successful companies .
lazear ( 2004 and 2005 ) , on the other hand , suggests that the successful entrepreneurs should be generalists a theory supported by astebro and thompson ( 2011 ) , who show that entrepreneurs tend to have a wider range of experiences than wage workers .
tetlock ( 1998 ) finds that generalists tend to be better forecasters than specialists .
in contrast , a wide variety of medical studies ( see , for example , hillner et al ( 2000 ) and nallamothu et al ( 2006 ) ) , show that outcomes tend to be better when patients are seen by specialists , rather than general practitioners .
however , none of this work considers the decision that individuals make with respect to being a specialist or generalist .
while some people will always become generalists due to personal taste , the question remains : is it ever rational to do so in the absence of a preference for interdisciplinarity ? and is the decision that the individual makes optimal from a societal perspective ?
there is evidence that being a generalist is costly .
adamic et al ( 2010 ) show that in a wide variety of contexts , including academic research , patents , and contributions to wikipedia , the contributions of individuals with greater focus tend to have greater impact , indicating that there is a tradeoff between the number of fields an individual can master , and her depth of knowledge in each .
this should not be surprising .
each of us has a limited capacity for learning new things by focusing on a narrow field of study , specialists are able to concentrate their efforts and maximize the use of that limited capacity , while generalists are forced to spread themselves more thinly in the pursuit of a wider range of knowledge . in the language of economics ,
generalists pay a fixed cost for each new field of study they pursue , in the form of effort expended learning new jargon , establishing new social contacts in a field , and becoming familiar with new literatures .
given that it is costly to diversify ones skills , the decision to become a generalist can be difficult to rationalize . in this paper ,
i examine model in which workers decide whether to be specialists or generalists to explore conditions under which it is rational for an individual to choose to be a generalist .
i show that when problems are single - dimensional and there are no barriers to working on problems in other disciplines , the equilibrium population contains only specialists .
however , when there are barriers to working on problems in other fields ( eg : communication barriers or institutional barriers ) then there is a tradeoff between the depth of study of the specialist and the wide scope of problems that the generalist can work on . when problems are relatively simple , generalists dominate because their breadth of experience gives them a wider variety of problems to work on .
but as problems become more difficult , depth wins out over scope , and workers tend to specialize .
i then show that the equilibrium decisions reached by individuals are not necessarily socially optimal .
as problems become harder , individual workers are more likely to specialize , but as a society , we would prefer that some individuals remain generalists .
this disconnect reflects the fact that from a social perspective , we would prefer to have researchers apply the widest possible variety of skills to the problems we face , but individuals internalize the cost of obtaining those skills .
thus , the model predicts that some populations will suffer from an undersupply of generalists . in such populations
, it would be socially beneficial to subsidize the acquisition of skills in broad subject areas .
finally , i consider an extension of the model in which problems have multiple parts .
this allows me to consider problems that are explicitly multidisciplinary that is , when different parts of a problem are best addressed using skills from different disciplines .
i show when problems are multidisciplinary , it is possible to rationalize being a generalist , even when there are no disciplinary boundaries . in particular , when there is a large advantage to applying the best tool for the job , being a generalist is optimal .
i construct a two period model . in period 1 , the workers face a distribution of problems and each worker chooses a set of skills . in period 2 ,
a problem is drawn from the distribution , and the workers attempt to solve it using the skills they acquired in period 1 .
i will solve for the equilibrium choice of skills in period 1 .
let @xmath0 be the set of all possible skills .
the skills are arranged into 2 disciplines , @xmath1 and @xmath2 , each with @xmath3 skills , @xmath4 .
an example with six skills arranged into two disciplines is shown in figure [ fig : two - disciplines ] .
two disciplines , each with three skills . ] a _ specialist _ is a person who chooses skills within a single discipline .
a _ generalist _ is a person who chooses some skills from both disciplines .
a problem , @xmath5 , is a task faced by the workers in the model .
a _ skill _ is a piece of knowledge that can be applied to the problem in an attempt to solve it .
each skill @xmath6 has either a high probability @xmath7 or a low probability @xmath8 of solving the problem .
_ _ i will define a problem by the matrix of probabilities that each skill will solve the problem .
that is , @xmath9 $ ] where @xmath10 if skill @xmath11 in discipline @xmath12 has a high probability of solving the problem and @xmath13 if it has a low probability of solving the problem .
so , for example , if there are two disciplines , each with three skills , a problem might be @xmath14\ ] ] meaning that two of the skills in each discipline have a high probability of solving the problem , and one skill in each discipline has a low probability of solving the problem .
define @xmath15 and @xmath16 .
the mechanics of the model are as follows . in period 1
, the workers , @xmath17 , each choose a set of skills @xmath18 . in period 2 , the workers attempt to solve a problem using those skills .
i will assume that workers have a _ capacity _ for learning skills , which limits the number of skills they can obtain . in the current context , i will assume that all workers all have the same capacity for learning new skills , and that all skills are equally costly to obtain .
let @xmath19 represent an individual s capacity for new skills and let @xmath20 be the cost of acquiring a new skill .
i assume that workers pay a fixed cost , @xmath21 , for learning skills in a new discipline .
that is , a worker pays @xmath22 to obtain the first skill in a discipline , and @xmath20 for every additional skill in that discipline . for simplicity
, i will assume that @xmath23 .
this assumption means that a specialist can obtain all @xmath3 skills in one discipline , and a generalist can obtain a total of @xmath24 skills spread over the two disciplines .
although workers in period 1 do not know the particular problem they will face in period 2 , they do know the distribution , @xmath25 , from which those problems will be drawn .
in particular , they know the probability that each skill will be an h skill or an l skill .
for simplicity , i will make two assumptions about the distribution of problems : 1 ) skills are _ independent , _ meaning that the probability that skill @xmath26 is an h skill is independent of the probability that skill @xmath27 is an h skill and 2 ) skills are _ symmetric within disciplines _ , meaning that every skill in a discipline has an equal probability of being an @xmath28 skill .
this knowledge of the distribution of problems can be translated into knowledge about individual skills .
let @xmath29 be the probability that a skill in discipline @xmath12 is an @xmath28 skill that is , @xmath30 $ ] where the expectation is taken over the distribution of problems , @xmath25 .
the vector of probabilities in the two disciplines , @xmath31 $ ] , is known _ ex ante . _
workers choose their skills in period 1 to maximize their expected probability of solving the problem in period 2 .
a nash equilibrium of this game is a choice of skill set for each worker in the population , @xmath32 such that no worker has an incentive to unilaterally change her skill set , given the distribution of problems .
in this section , i consider two questions .
the first question concerns individual decision - making what is the equilibrium skill acquisition decision of the workers ? under what conditions do individuals decide to generalize ?
the second question concerns the optimality of that population from an organizational perspective . is the equilibrium population optimal ?
note that in order to simplify the exposition , i will consider a special case where all disciplines are equally useful in expectation that is , where @xmath33 .
it is straightforward to generalize the results to a case where @xmath34 ( see appendix for the details ) . given that generalists pay a significant penalty for diversifying their skills , it is difficult to explain the existence of generalists in the population .
theorem [ thm : if open no generalists ] states that if workers can work on any available problem , then there will be no generalists in equilibrium .
[ thm : if open no generalists]if skills are independent and symmetric within discipline , and workers can work on any available problem , then no worker will ever want to be a generalist and the equilibrium population will contain only specialists . the _ ex ante _ probability that a specialist in discipline @xmath35 will be able to solve a problem from a given distribution , @xmath25 , is @xmath36 & = & \sum_{y}prob\left(\mbox{one of skills solves } y\right)*\delta\left(y\right)\\ & = & \sum_{y}\left(1-prob\left(\mbox{none do}\right)\right)*\delta\left(y\right)\\ & = & 1-\sum_{n_{i}=0}^{k}h^{n_{i}}l^{k - n_{i}}\left({k\atop n_{i}}\right)\delta^{n_{i}}\left(1-\delta\right)^{k - n_{i}}\\ & = & 1-\left(\delta h+\left(1-\delta\right)l\right)^{k}\end{aligned}\ ] ] where @xmath37 is the number of @xmath28 skills in discipline @xmath35 in a particular problem , @xmath5 . now , consider a generalist who is spreading his skills across both disciplines . the _ ex ante _ probability that a generalist with @xmath38 skills in discipline @xmath39 and @xmath40 skills in discipline 2 will solve a problem from a given distribution , @xmath25 , is @xmath41 & = & 1-\sum_{y}prob\left(\mbox{none of skills solve } y\right)*\delta\left(y\right)\\ & = & 1-\left(\delta h+\left(1-\delta\right)l\right)^{k - c}\end{aligned}\ ] ] @xmath42 , and
thus no individual will ever be a generalist in two disciplines .
( see appendix for the same result with @xmath34 ) note that this result generalizes to a case with more than two disciplines .
generalists do worse as they add skills in additional disciplines , so this result holds regardless of the number of disciplines a generalist spreads himself across .
theorem [ thm : if open no generalists ] clearly indicates that when workers can solve problems in other fields , there is no advantage to being a generalist .
however , in practice , there may be many barriers between disciplines that prevent a worker in one discipline from solving problems in another .
cultural or institutional barriers may prevent her from working on questions in other disciplines , either because resources are not forthcoming or because it is difficult to get compensated for work in other areas .
communication barriers are also a significant impediment to interdisciplinary work
although a software engineers may have skills useful in solving user interface problems , field - specific jargon may make it difficult for her to communicate her insights .
if communication barriers are severe enough , she may even have difficulty understanding what open questions exist .
finally , a person in one field may simply be unaware of problems that exist in other fields , even if her skills would be useful in solving them .
barriers to working on problems outside ones discipline give us the ability to talk about the `` scope '' of a worker s inquiry .
generalists are able to work on a broader set of problems , and thus their scope is larger than that of specialists .
there is therefore a tradeoff between the depth of skill gained through specialization and the scope gained through generalization .
a specialist has a depth of skill that gives her a good chance of solving the limited set of problems in the area she specializes in .
generalists have a limited number of skills , but are able to apply those skills to a much broader set of problems .
thus , the choice between being a specialist and a generalist can be framed in terms of a tradeoff between the depth of one s skill set and scope of one s problem set . more formally ,
choice of whether to specialize will depend on two parameters .
first , let @xmath43 be the expected probability that a skill wo nt be able to solve a problem drawn from @xmath25 .
when @xmath44 is large , the probability that any one skill will solve the problem is very low .
thus , we can think of problems becoming more difficult as @xmath44 increases .
second , let @xmath45 be the fraction of all problems that occur in discipline 1 . when @xmath45 is very large or very small , most of the problems fall in one field or another , limiting the value of increasing the scope of the problem set . these two parameters@xmath45 and @xmath44define a range in which workers will choose to generalize in equilibrium .
this range is illustrated in figure [ fig : generalists range ] .
this diagram illustrates the tradeoff between depth and scope .
when problems are easy to solve , scope is more valuable than depth .
however , as @xmath44 increases and problems become more difficult , depth wins out over scope , and the range in which individuals choose to generalize shrinks .
equilibrium skill acquisition decisions when @xmath46 and @xmath47 . ]
theorem [ thm : communication barriers and generalists ] summarizes these results .
[ thm : communication barriers and generalists]if skills are independent and symmetric within discipline , and there are barriers to working on problems in other disciplines , then workers will generalize if @xmath48 where @xmath45 is the fraction of problems assigned to discipline 1 .
if @xmath49 , then workers will all specialize in discipline 1 and if @xmath50 then workers will all specialize in discipline 2 . in this case , the _ ex ante _ probability that a problem is solved by a specialist is @xmath51 for a specialist in discipline 1 and @xmath52 for a specialist in discipline 2 .
since generalists can work on problems in both disciplines , their expected probability of solving the problem is @xmath53 .
a worker will generalize if @xmath54<e\left[p\left(g\right)\right]$ ] and @xmath55<e\left[p\left(g\right)\right]$ ] .
the result follows immediately .
( see appendix for the same result with @xmath34 ) note that the size of the regions in which workers specialize depends on how costly it is to diversify ones skills . as the fixed cost of learning something in a new discipline increases ( @xmath56 )
, the regions in which people specialize grow . in this section ,
i consider whether this distribution of specialists and generalists in the population is optimal , from a societal perspective .
there is reason to believe that it would not be . from a societal standpoint
, we would like to maximize the probability that someone manages to solve the problem .
this means that as a society , we would prefer to have problem solvers apply as wide a range of skills as possible . but workers who diversify their skills obtain fewer skills overall , which tends to make the individual want to specialize .
the result of this disconnect between individual and social welfare is a range in which generalists are under provided ( see figure [ fig : regions of social suboptimality ] ) .
as problems become more difficult , this region of suboptimality grows , as is illustrated in figure [ fig : regions of social suboptimality grow ] . ]
regions of social suboptimality for @xmath46 , @xmath57@xmath47 , @xmath58 ] theorem [ thm : optimality of the equilibrium ] summarizes these results .
[ thm : optimality of the equilibrium]if skills are independent and symmetric within discipline , and there are barriers to working on problems in other disciplines , then there is a range of values for @xmath45 ( the fraction of problems assigned to discipline 1 ) such that generalists are underprovided in the equilibrium population of problem solvers .
in particular , generalists are underprovided when @xmath59 or @xmath60 .
the probability that at least one of the @xmath61 problem - solvers in the population solves the problem is @xmath62 .
if all of the individuals in the population are specialists in discipline 1 , then with probability @xmath45 , each specialist has a probability @xmath63 of solving the problem and @xmath64 of not solving it . with probability @xmath65
, the problem is assigned to the other discipline , and no specialist solves it .
thus , the probability of someone in a population of discipline 1 specialists solving the problem is @xmath66\\ & = & 1-\left[\phi prob\left(\mbox{one fails}\right)^{n}+\left(1-\phi\right)*1\right]\\ & = & 1-\left[\phi\left(\pi^{k}\right)^{n}+\left(1-\phi\right)*1\right]\\ & = & \phi\left(1-\pi^{kn}\right)\end{aligned}\ ] ] on the other hand , if they are all generalists , then the probability of at least one solving the problem is @xmath67 society is better off with a population of generalists when @xmath68 , which is true when @xmath69 .
however , there is a population of generalists when @xmath70 .
it is always the case that @xmath71 .
so if @xmath59 , then society is better off with a population of generalists , but has a population of specialists .
we can make a similar argument for specialists in discipline 2 .
society is better off with a population of generalists when @xmath72 , which is true when @xmath73 .
however , there is a population of generalists when @xmath74 .
it is always the case that @xmath75 .
so if @xmath60 , then society is better off with a population of generalists , but has a population of specialists .
( see appendix for the same result with @xmath34 ) note that the size of the regions of suboptimality will depend on the number of individuals in the population . as @xmath61 increases ,
the suboptimal regions become larger .
in the previous section , i showed that barriers to addressing problems in other disciplines can induce problem solvers to diversify their skills . in this section ,
i consider an extension of the previous model , which highlights a second scenario in which individuals can be incentivized to acquire skills in multiple disciplines : problems with multiple parts .
as problems become increasingly complicated , they may be broken down into many different sub - problems .
although in some cases , these subproblems may all be best addressed within a single discipline , in others , different subproblems will be best addressed using different skills . in this section ,
i show that when problems are _ multidisciplinary_that is , when different parts of a problem are best addressed using different disciplines then a population of generalists can be sustained . as in the previous model ,
skills in the set @xmath0 are divided into two disciplines , @xmath1 and @xmath2 .
workers use their skills to address a problem , the nature of which is not known _ ex ante . _
they will choose to be a specialist or generalist in period 1 to maximize their chances of solving the problem in period 2 .
_ _ but now , suppose each problem consists of two parts , @xmath76 and @xmath77 . in order to solve the problem
, an individual must solve all parts of the problem .
each part of the problem is addressed independently by the skills in each of the disciplines .
thus , much as before , we can define the parts of the problem by a matrix of probabilities that each skill will solve the problem .
that is , @xmath78 $ ] where @xmath79 if skill @xmath11 in discipline @xmath12 has a high probability of solving part @xmath35 and @xmath13 if it has a low probability of solving part @xmath35 . as in the previous section
, i will assume that for each part of the problem , skills are independent _ _
( @xmath80 uncorrelated with @xmath81 ) and skills are symmetric within disciplines ( @xmath82 ) . as before , the probability that a given skill is an @xmath28 skill
is not known _ ex ante .
_ however , the workers know the expected probability that a skill is an @xmath28 skill
. i will allow the expected probabilities to vary across parts of the problem in other words , it is possible that a discipline will be more useful in solving one of the parts of the problem than in solving the other part of the problem .
let @xmath83 be the probability that a skill from discipline @xmath12 is an h skill for part @xmath35 of the problem .
that is , @xmath84 $ ] .
the matrix @xmath85 $ ] describes a distribution of problems , @xmath25 , and is known _ ex ante . _
an entry in the @xmath86 column of that matrix is the vector of probabilities that a skill in each of the disciplines will be useful for solving part @xmath35 of the problem .
we can categorize the problems according to the relative usefulness of the two disciplines in the two parts of the problem .
there are two categories for the problems : 1 .
one discipline is as or more useful for both parts of the problem : @xmath87 2 .
one discipline is more useful for part 1 and the other discipline is more useful for part 2 : @xmath88 and @xmath89 these categories are illustrated in figure [ fig : taxonomy of distributions ] . ]
if a problem falls into the first category , then the results are similar to those obtained in section [ sec : specialization and barriers between disciplines ] .
in particular , if there are no barriers to working on problems in other disciplines , then all workers will specialize .
if a problem falls into the second category , then the results do not resemble any of those already explored .
problems with multiple parts , each of which is best addressed within the context of a different discipline , are often referred to as _ multidisciplinary .
_ generalists have an advantage in multidisciplinary problems , because they can apply different types of skills to different parts of a problem .
for example , suppose a scientist is look at nerve conduction in an organism .
that problem may have elements are are best addressed using biological tools , and other elements that are best addressed using physics tools .
an individual with both biology and physics skills will have an advantage over someone who is forced to use ( for example ) physics skills to solve both parts of the problem .
the below states that when problems are multidisciplinary , it can be rational to be a generalist , even in the absence of barriers to working in other fields . more formally ,
suppose that if a worker uses the `` right '' discipline for a part of a problem , then there is a probability @xmath90 that a skill in that discipline is useful ( @xmath91 when @xmath12 is the right discipline to use for part @xmath35 of the problem ) .
if she uses the `` wrong '' discipline , then there is a probability @xmath92 that a skill in that discipline is useful ( @xmath91 when @xmath12 is the wrong discipline to use for part @xmath35 of the problem ) .
this is without loss of generality , because the only thing that makes a problem multidisciplinary is the ordering of the usefulness of the disciplines .
further , let @xmath93 and @xmath94 .
these represent the probability that a skill in the right discipline will not solve a part of a problem and the probability that a skill in the wrong discipline will not solve part of the problem .
note that @xmath95 .
when the efficacy of the two disciplines is very different ( @xmath96 ) , then using the right skill for the job has a large effect on the probability of solving the problem as a whole , and it will be rational to obtain skills in multiple disciplines .
figure [ fig : regions of specialization in two part problem ] illustrates the region in which individuals choose to be generalists and specialists , and theorem [ thm : multidisciplinary problems ] summarizes the result .
equilibrium skill acquisition decisions when problems are multidisciplinary , and @xmath46 and @xmath47 . ]
[ thm : multidisciplinary problems]if skills are independent and symmetric within discipline , and problems multidisciplinary ( eg : @xmath97 $ ] with @xmath98 ) then there is a set of values of @xmath93 and @xmath94 such that it is individually optimal for workers to be generalists , even when there are no barriers to solving problems in other fields . in particular , workers will become generalists when @xmath99 wlog , consider the case where @xmath97 $ ] with @xmath98 . a specialist in discipline
@xmath35 will have @xmath3 skills in discipline @xmath35 .
the expected probability that her skills will solve _ both _ parts of the problem is @xmath100*e\left[p\left(\mbox{success on part 2}\right)\right]$ ] , which is @xmath101 where @xmath93 and @xmath94 .
a generalist will have skills in both disciplines . in this case
, it will be optimal for a generalist to split her skills evenly between the two disciplines , and she will obtain @xmath102 skills in each . the expected probability that she will solve both parts of the problem is @xmath103 .
thus , individuals choose to generalize , when @xmath99 .
, we have a case that fits into the first category in the taxonomy of problem distributions in figure [ fig : taxonomy of distributions ] .
since @xmath104 , we can use this calculation to verify the claim made above that in the case where skills are symmetric , the results are the same as in section [ sec : specialization and barriers between disciplines ] . ] this region , as a function of @xmath105 and @xmath106 , is illustrated in figure [ fig : regions of specialization in two part problem ] for @xmath107 and @xmath47 .
the boundary of this region is defined by the equation @xmath108 . as would be expected ,
the region where individuals specialize shrinks as the costs to generalizing ( @xmath21 ) become smaller , relative to the individual s total capacity for learning new skills @xmath109 .
being a generalist is costly .
every new area of expertise comes at considerable fixed cost , in the form of a new literature , new jargon , and new basic ideas
. however , there are clearly a large ( and growing ) number of individuals in research communities who choose to do so .
this raises the question of whether that decision is ever individually rational ? and is there a reason to believe that fewer people choose to be generalists than is socially optimal ?
this paper suggests that being a generalist can be a rational decision under particular conditions . in particular
, obtaining a broad range of skills is rational if there are significant barriers to working on questions in fields with which one is unfamiliar . those who pay the initial price of learning the jargon and literature of a new field
reap the benefits in the form of a larger pool of problems to solve .
it can also be rational to be a generalist if problems are multidisciplinary that is , if different parts of a problem are best addressed using skills in different disciplines .
moreover , because individuals bear the costs of becoming generalists , we will tend to have fewer of them than is optimal from a societal standpoint .
this potential market failure means that in some cases , it is optimal for funding agencies and private organizations to subsidize individuals in their efforts to diversify their skills and promote interdisciplinary researchers .
however , it is unclear whether our current situation is one in which such funding is required .
more careful consideration of this question is a good candidate for further work .
there are several elements of this model that suggest directions for future research
. it would be interesting to consider a case where individuals differ in their innate capacity for learning skills .
this might provide some insight into what types of individuals choose to become generalists .
incorporating collaboration would be another particularly interesting extension .
collaboration has always been an important part of problem solving and innovation , and it has only become more important over time ( see , among others , laband and tollison ( 2000 ) , acedo et al ( 2006 ) , and goyal et al ( 2006 ) ) .
there is reason to believe that in a collaborative context , the advantage to generalists would be enhanced , because generalists could connect specialists in different fields . on a more general level , there is much to be gained from a better understanding of specialization decisions .
research universities , government organizations such as nasa , and private enterprises ranging from genentec to google are reliant on the skills of individual problem solvers .
the decisions these individuals make about the breadth skills they obtain have an undeniable effect on the rate of innovation .
however , we still have only a limited understanding of the what drives those skill acquisition decisions , and what distinguishes the role of specialists and generalists in problem solving .
better theoretical models of these decisions have the potential to greatly enhance our understanding of this important aspect of such organizations .
references acedo , f .. j . , barroso , c. , casanueva , c. , and galan , j.l . , 2006 .
co - authorship in management and organizational studies : an empirical and network analysis .
journal of management studies 43 ( 5 ) _ , _ 957 - 983 adamic , l.a . , wei , x. , yang , j. , gerrish , s. , nam , k.k . , and clarkson , g.s . , 2010 .
individual focus and knowledge contribution .
arxiv:1002.0561v1 astebro , t. , and thompson , p. , 2011
entrepreneurs , jacks of all trades or hobos ?
research policy 40 ( 5 ) , 637 - 649 berlin , i. , 1953 . the hedgehog and the fox .
simon and schuster , new york .
collins , j. , 2001 .
good to great .
harpercollins , new york .
goyal , s. , van der leij , m. j. , and moraga - gonzalez , j.l . , 2006 .
economics : an emerging small world__. _ _ journal of political economy 144 ( 2 ) _ , _ 403 - 412 hillner , b.e . , smith , t.j . , and desch , c.e . ,
2000 . hospital and physician volume or specialization and outcomes in cancer treatment : importance in quality of cancer care .
journal of clinical oncology 18 ( 11 ) , 2327 - 2340 kremer , m. , 1993 .
the o - ring theory of economic development .
quarterly journal of economics 108 ( 3 ) , 551 - 575 laband , d.n . , and tollison , r.d . , 2000 .
intellectual collaboration .
journal of political economy 108 ( 3 ) , 632 - 662 lazear , e. , 2004 .
balanced skills and entrepreneurship .
american economic review 94 ( 2 ) , 208211 .
lazear , e. , 2005 .
journal of labor economics 23 ( 4 ) , 649680 .
nallamothu , b.k , wang , y. , magid , d.j . ,
mcnamara , r.l . ,
herrin , j. , bradley , e.h . ,
bates , e.r . ,
pollack , c.v . , and krumholz , h.m . , 2006 .
relation between hospital specialization with primary percutaneous coronary intervention and clinical outcomes in st - segment elevation myocardial infarction : national registry of myocardial infarction-4 analysis .
circulation 113 , 222 - 229 strober , m.h .
habits of the mind : challenges for multidisciplinary engagement .
social epistemology 20 ( 3 ) , 315 - 331 tetlock , p.e .
close - call counterfactuals and belief - system defenses : i was not almost wrong but i was almost right .
journal of personality and social psychology 75 ( 3 ) , 639 - 652
theorem [ thm : if open no generalists-1 ] is the equivalent of theorem [ thm : if open no generalists ] , and states that if individuals can work on any available problem , then there is no advantage to being a generalist .
[ thm : if open no generalists-1]if skills are independent and symmetric within discipline , and workers can work on any available problem , then no worker will ever want to be a generalist and the equilibrium population will contain only specialists . as above , the _ ex ante _ probability that a specialist in discipline @xmath35 will be able to solve a problem from a given distribution , @xmath25 , is @xmath36 & = & 1-\left(\delta_{i}h+\left(1-\delta_{i}\right)l\right)^{k}\\ & = & 1-\pi_{i}^{k}\end{aligned}\ ] ] where @xmath110 .
wlog , suppose @xmath111 . since @xmath112
, this means that @xmath113 and @xmath54>e\left[p\left(s_{2}\right)\right]$ ] .
thus , to determine whether any individual will generalize , i need to compare @xmath54 $ ] to @xmath114 $ ] .
the _ ex ante _ probability that a generalist with @xmath38 skills in discipline 1 , and @xmath40 skills in discipline 2 solves a problem from a given distribution , @xmath25 , is @xmath41 & = & 1-\left(\delta_{1}h+\left(1-\delta_{1}\right)l\right)^{x}\left(\delta_{2}h+\left(1-\delta_{2}\right)l\right)^{k - c - x}\\ & = & 1-\pi_{1}^{x}\pi_{2}^{k - c - x}\end{aligned}\ ] ] since @xmath113 , @xmath114 $ ] is strictly increasing in @xmath38 .
this means that a generalist will set @xmath115 and @xmath114=1-\pi_{1}^{k - c-1}\pi_{2}$ ] , which is clearly less than @xmath54=1-\pi_{1}^{k}$ ] .
theorem [ thm : communication barriers and generalists-1 ] is a generalized version of theorem [ thm : communication barriers and generalists ] , and states the parameter range in which individuals will choose to diversify their skills when there are barriers to working interdisciplinarily .
[ thm : communication barriers and generalists-1]if skills are independent and symmetric within discipline , and there are barriers to working on problems in other disciplines , then there is a range of values for @xmath45 ( the fraction of problems assigned to discipline 1 ) for which individuals will generalize .
if @xmath111 , then workers will obtain @xmath118 skills in discipline 1 and one skill in discipline 2 when @xmath119 , @xmath3 skills in discipline 1 when @xmath120 , and @xmath3 skills in discipline 2 when @xmath121 .
if @xmath122 , then workers will obtain @xmath118 skills in discipline 2 and one skill in discipline 1 when @xmath123 , @xmath3 skills in discipline 1 when @xmath124 , and @xmath3 skills in discipline 2 when @xmath125 . in this case ,
the _ ex ante _ probability that a problem is solved by a specialist is @xmath126 for a specialist in discipline 1 and @xmath127 for a specialist in discipline 2 .
since generalists can work on problems in both disciplines , their expected probability of solving the problem is @xmath128 where @xmath38 is the number of skills the generalist chooses to acquire in discipline 1 .
first , suppose @xmath111 .
since @xmath112 , this means that @xmath113 and @xmath114 $ ] is strictly increasing in @xmath38 .
thus , a generalist will choose a minimal number of skills in the less useful discipline , and @xmath114=1-\pi_{1}^{k - c-1}\pi_{2}$ ] .
an individual will generalize if @xmath54<e\left[p\left(g\right)\right]$ ] and @xmath55<e\left[p\left(g\right)\right]$ ] . setting
@xmath129 implies that @xmath130 .
setting @xmath131 implies that @xmath132 .
we can verify that in the appropriate ranges , individuals choose to specialize .
the result follows immediately .
the proof for @xmath122 is similar . for the proof when @xmath133 , see theorem [ thm : communication barriers and generalists ] . finally , theorem [ thm : optimality of the equilibrium-1 ] is the generalization of theorem [ thm : optimality of the equilibrium ] .
it states that there is a parameter region in which individuals choose to specialize , but society would prefer to have at least a few generalists .
[ thm : optimality of the equilibrium-1]if skills are independent and symmetric within discipline , and there are barriers to working on problems in other disciplines , then there is a range of values for @xmath45 ( the fraction of problems assigned to discipline 1 ) such that generalists are underprovided in the equilibrium population of problem solvers . first , suppose that @xmath111 . the probability that at least one of the @xmath61 problem - solvers in the population solves the problem is @xmath62 .
if all of the individuals in the population are specialists in discipline 1 , then every individual has probability @xmath45 of a problem occurring in her discipline . in that case
, each specialist in discipline has a probability @xmath138 of solving the problem and @xmath139 of not solving it . with probability @xmath65
, the problem is assigned to the other discipline , and no specialist solves it .
thus , the probability of someone in a population of specialists solving the problem is @xmath140\\ & = & 1-\left[\phi prob\left(\mbox{one fails}\right)^{n}+\left(1-\phi\right)*1\right]\\ & = & 1-\left[\phi\left(\pi_{1}^{k}\right)^{n}+\left(1-\phi\right)*1\right]\\ & = & \phi\left(1-\pi_{1}^{kn}\right)\end{aligned}\ ] ] society is better off with a population of generalists than a population of discipline 1 specialists when @xmath143 , which is true when @xmath144 .
however , there is a population of generalists when @xmath130 .
it is always the case that @xmath145 .
thus , if @xmath134 , then society is better off with a population of generalists , but has a population of specialists . through a similar argument ,
society is better off with a population of generalists than a population of discipline 2 specialists when @xmath146 , which is true when @xmath147 .
however , there is a population of generalists when @xmath148 .
it is always the case that @xmath149 .
thus , if @xmath135 , then society is better off with a population of generalists , but has a population of specialists . | many organizations rely on the skills of innovative individuals to create value , including academic and government institutions , think tanks , and knowledge - based firms . roughly speaking
, workers in these fields can be divided into two categories : specialists , who have a deep knowledge of a single area , and generalists , who have knowledge in a wide variety of areas . in this paper ,
i examine an individual s choice to be a specialist or generalist .
my model addresses two questions : first , under what conditions does it make sense for an individual to acquire skills in multiple areas , and second , are the decisions made by individuals optimal from an organizational perspective ?
i find that when problems are single - dimensional , and disciplinary boundaries are open , all workers will specialize .
however , when there are barriers to working on problems in other fields , then there is a tradeoff between the depth of the specialist and the wider scope of problems the generalist has available .
when problems are simple , having a wide variety of problems makes it is rational to be a generalist .
as these problems become more difficult , though , depth wins out over scope , and workers again tend to specialize .
however , that decision is not necessarily socially optimal on a societal level , we would prefer that some workers remain generalists . skill acquisition , specialization , jack - of - all - trades , problem solving , knowledge based production , human capital _ jel codes : j24 , o31 , d00 , m53 , i23 _ _ thanks to scott page and ross oconnell .
this work was supported by the nsf and the rackham graduate school , university of michigan .
computing resources supplied by the center for the study of complex systems , university of michigan . |
nowadays the study of the exotic hadrons is one of the hottest topics of hadron physics . within the quark model
the ordinary mesons are considered as bound states of quark - antiquark pairs and baryons are three - quark bound states .
however , more complicated states which include additional quarks and/or valence gluons are not forbidden by qcd .
there were three big waves in the history of the exotic hadrons .
the first wave was created by the pioneering papers by jaffe @xcite on spectroscopy of the four quark @xmath2 states within mit bag model with light @xmath3 quarks .
these states we call now light tetraquarks .
the second wave was created by the prediction of the @xmath4 strange pentaquark state with very small width within the chiral soliton model @xcite .
the third and last wave was generated by the belle collaboration discovery of the @xmath5 meson with very small width and the mass near @xmath6 threshold @xcite .
now we have many candidates for exotic tetraquark states with heavy quark content which are called the @xmath0 mesons .
there are also several candidates for the heavy pentaquarks ( see , for example , the recent reviews @xcite ) .
a lot of the different microscopical models have been suggested to explain their unusual properties .
we discuss a new approach to the @xmath0 mesons based on the possibility of a large mixing between radial excitations in the light quark and antiquark system and states with heavy quark content .
such mixing might be induced by the nonperturbative quark - gluon interaction related to the nontrivial topological structure of the qcd vacuum .
the second part of this review is devoted to the modern status of glueballs which are the bound states of the gluons .
we discuss the properties of these exotic hadrons including the possibility of the existence of the three - gluon bound states . and
finally , we show that glueballs can influence strongly the properties of quark - gluon plasma .
one of the main problems of the exotic hadron spectroscopy is the possibility of large mixings between exotic and nonexotic hadrons with the same quantum numbers .
this mixing does not allow , in the most cases , to separate exotic states from the ordinary hadrons .
the instantons , the strong fluctuations of the vacuum gluon fields , induce thooft s multiquark interaction which can lead to the strong mixing of the different quark flavors ( see review @xcite ) . due to the specific structure of thooft s interaction
it can violate the ozi rule in the channels with the quantum numbers @xmath7 and @xmath8 as shown in fig.1a .
since the instanton corresponds to a subbarrier transition between vacua with the different topological charge , the natural energy scale of this violation is related to the height of the potential barrier between these vacua .
this height is given by the energy of the so - called sphaleron @xmath9 gev with @xmath10 fm being the average instanton size in the qcd vacuum @xcite .
for higher mass excitations one can expect that the instanton - anti - instanton molecules give an important contribution to the mixing between states with different quark flavors .
examples of such mixing between light quark - antiquark states and states with heavy quarks are shown in the figs.1b , c and fig.2 . the nonperturbative interaction which governs that mixing is the quark - gluon effective interaction induced by instantons @xcite @xmath11 where @xmath12 is the gluon field strength
, @xmath13 is the effective mass of the light @xmath14 quarks in the vacuum and @xmath15 is the anomalous quark chromomagnetic moment ( aqcm ) .
an extension of eq.[lag ] which preserves the chiral symmetry was suggested by polyakov and has the following form @xcite @xmath16 where @xmath17 is the pion decay constant . case . ] within the instanton model , the value of aqcm is @xcite @xmath18 for @xmath19 mev in mean field approximation .
in general , a high radially excited state of a light quark - antiquark system can mix easily with states which have a heavy quark content , if they have the same quantum numbers and approximately the same mass .
it might be that the @xmath5 and the @xmath20 are result of the mixing of molecular states and/or tetraquark charmed hidden states with high radial excitations of a light quark system , by the mechanism shown in fig.1b , c and fig.3b .
a recent paper in the spirit of this idea was presented by coito @xcite . in the case of such mixing one
should observe the decays of the xyz mesons to final states without heavy quarks .
the absence of these decay modes in present experiments might be related to the existence of many nodes in the wave function of the light quark system .
one can estimate , for example , that light quark @xmath21 systems with radial numbers @xmath22 have approximately the same mass as the @xmath5 and the @xmath20 .
the many node structure of the wave function might lead to a very small overlap of the initial and final wave functions in the decay .
these strong effect of the many nodes in wave function in decays is well known phenomenon .
in particular , brodsky and karliner in @xcite used it to explain the small rate of @xmath23 in comparison with @xmath24 .
therefore , the mixing of high excited states in light quark - antiquark systems with heavy quark states might be behind of the unusual properties of xyz mesons .
there is a longstanding problem to explain the absence of transverse polarization of @xmath1 in the inclusive production at high energy .
the transverse polarization was predicted by perturbative qcd and it comes from the fragmentation of on - shell gluon to charm - anticharm pair ( see review @xcite ) .
such polarization is in disagreement with the measurement at tevatron @xcite and at lhc @xcite . in a recent paper the polarization of gluons in the constituent quark induced by instantons
was calculated @xcite .
the diagram without pion , fig.4a , produces no polarization of the gluons @xcite .
however , the diagram with an additional pion , fig.4b , leads to polarization of the gluons . in this case
the instanton and anti - instanton provide a polarization of opposite sign .
the contribution of the instantons to the @xmath1 polarization in proton - proton collisions is presented in fig.5 .
it is evident , that such mechanism leads to a longitudinal polarization of the @xmath1 because the perturbative vertex with a hard gluon does not change the helicity of the charm quark .
-0.5 cm from the another side , one can consider this new mechanism as the contribution to the inclusive production of the quarkonia coming from the mixing of light quarks with a state having heavy quark content .
the glueballs , the bound states of the gluons , are now widely under discussion ( see the reviews @xcite ) .
different approaches including mit bag model , several types of the constituent models , holographic qcd , qcd sum rule methods and the lattice simulation were used to extract the information on the spectrum and quantum numbers of the glueballs .
one direct way based on the first principals of qcd to calculate the glueball spectrum are lattice calculations . at the present , the old @xcite and more recent lattice calculations @xcite in quenched qcd , i.e. the theory without quarks
, are in the agreement and the result for the low mass of the glueballs in pure @xmath25 is @xmath26 one of the interesting features of lattice calculations is that in the unquenched qcd with light quarks the result for the lowest mass of glueballs practically does not change @xcite @xmath27 for 2 + 1 flavor and @xmath28 mev .
a more recent result for @xmath29 and @xmath30 mev is @xcite @xmath31 there are many candidates for the glueball states . for the scalar glueballs we have the following mesons @xmath32 for pseudoscalar glueball the candidates are @xmath33 in the case of the tensor glueballs there are two candidates @xmath34 one of the main problems of the glueball spectroscopy
is the possible large mixings of the glueballs with ordinary and exotic quark states which leads to the difficulties in disentangling the glueballs in the experiments ( see recent discussion in @xcite ) .
the example of such mixing is presented in fig.6 .
the contribution of the mixing presented in fig.6a was considered in @xcite within the qcd sum rule method , and it was shown to be very important .
one of the suitable ways to avoid this problem and obtain a clear prediction for the glueballs is to study the glueballs with exotic quantum numbers . @xmath35 which can not mix with ordinary quark - antiquark states .
however , the calculation of the masses and decay modes of these exotic glueballs is very difficult because one should consider states with a large number of gluons .
for example , lattice calculations give a very large mass @xmath36 gev with a large uncertainty for the @xmath37 exotic three - gluon glueball .
the first attempt to calculate the masses of the glueballs with exotic quantum number was done in ref .
@xcite within the qcd sum rule approach and a rather small mass for @xmath38 glueball was found .
however in a recent paper @xcite , it was shown that the interpolation current for the gluonic state used in @xcite has an anomaly and can not be applied for the glueball calculations . in ref .
@xcite a new interpolating current for the @xmath38 glueball has been suggested .
the derived qcd sum rules are consistent and stable .
the resulting value for the mass of the @xmath38 glueball state is @xmath39 gev and an upper limit for the total width @xmath40 mev has been obtained .
it has been also argued that the mixing of this glueball state with @xmath38 tetraquark is expected to be very small .
therefore , the exotic @xmath38 glueball can be considered a pure gluon state . at the present ,
the belle collaboration is searching for this state @xcite .
in experiments at rhic and lhc a new type of nuclear matter the so - called strongly interacted quark - gluon plasma has been discovered ( see review @xcite ) .
it behaves as a liquid and does not have the expected gas - like behavior .
there are many experiments planned at the running and future facilities to investigate the properties of such new matter .
it is very important from the point of view of theory of the strong interactions to find the fundamental reason which leads to such behavior of quark - gluon matter .
one possible way is to investigatethe simpler case of pure gluodynamics , the theory without quarks .
recently very precise lattice results for the equation of state ( eos ) of this theory at finite @xmath41 below and above the deconfinement temperature @xmath42 were presented @xcite .
they challenge our understanding of qcd dynamics at finite temperatures .
one of the puzzles is a very spectacular behavior of the trace anomaly , @xmath43 , as a function of @xmath41 . in particular , just above
@xmath42 it rapidly grows untill @xmath44 and then it decays as @xmath45 untill @xmath46 .
a mechanism which can explain such anomalous behavior has been suggested in ref .
the mechanism is based on the possibility of a large change of the glueball masses above the @xmath42 .
the starting point is the relation between the lowest scalar glueball mass , @xmath47 , and the gluon condensate , @xmath48 at @xmath49 , which appears naturally in the dilaton approach @xcite @xmath50 where @xmath51 is the glueball coupling constant to gluons .
lattice calculations show that the gluon condensate decreases roughly by a factor two at @xmath52 due to the strong suppression of its electric component , while slightly above @xmath42 the condensate vanishes very rapidly due to the cancellation between its magnetic and electric components @xcite .
the temperature behavior of the condensate at @xmath53 can be described by the equation @xcite @xmath54 \nonumber\ ] ] and the mass of scalar and pseudoscalar glueballs for @xmath55 are given by @xmath56 the possible dynamical reason for the decrease of scalar and pseudoscalar glueball masses above @xmath42 is the strong attraction between gluons induced by instanton - antiinstanton molecules shown in fig.7 .
such a molecular structure of the instanton vacuum is expected above @xmath42 for the quark - antiquark channel the effect of the instanton - antiinstanton induced interaction was discussed in @xcite and it was shown to be strong . for the gluon - gluon interaction the effect should be even stronger because it is enhanced by factor @xmath57 , where @xmath58 is the instanton action .
the rapid decrease of the scalar and preudoscalar glueball masses above @xmath42 leads to a large contribution of these glueballs to the pressure of the gluon plasma and the sum of the glueball and free gluon contributions describe the lattice data very well as shown in fig.8 .
given these results we can describe three different phases of gluon matter @xcite .
the first phase is the confinement regime at @xmath59 and is described by a gas of massive glueballs .
the second phase appears just above @xmath42 at @xmath60 , where @xmath44 . in this phase
the main contribution to the eos is coming from the very light scalar and pseudoscalar glueballs . above the @xmath61
the third phase appears , a mixed phase of massless gluons and point - like scalar - pseudoscalar massless glueballs .
the existence of such phases can be checked in high multiplicity events in relativistic heavy ion collisions where an abundant production of glueballs is expected @xcite . the most simple way is to look for the change of the scalar and pseudoscalar glueball masses in the two pion and three pion channels , correspondingly , as a function of multiplicity .
finally it should be also mentioned that the mixing between quark and gluon states can change strongly above @xmath42 due to the transition between single instanton induced mixing to the mixing produced by instanton - antiinstanton molecules .
the exotic hadrons carry a very important information on the structure of the strong interactions and the properties of the exotic hadrons are very sensitive to the nonperturbative structure of qcd .
we have shown that the xyz mesons can arise as the result of a nonperturbative mixing between different configurations with and without heavy quark content .
we have furthermore suggested a new mechanism for heavy quarkonia production at high energy which can be related to such mixing .
we have also discussed the modern status of the glueballs and their mixing with quarkonia . finally , we have predicted a drastic change of the scalar and pseudoscalar glueball masses above the deconfinement temperature in the gluonic plasma which has very exciting consequences at the level of the possible phase transitions at very high temperatures .
i would like to thank alexander dorokhov , sergo gerasimov , mikhail ivanov , ernst - michael ilgenfritz , makoto oka , pengming zhang , bing - song zou and , especially , vicente vento for useful discussions .
i am grateful to ahmed ali and michail ivanov for the invitation to give a talk about exotic hadrons at helmholtz - dias international summer school `` quantum field theory at the limits : from strong fields to heavy quarks '' , 18 - 30 july 2016 , dubna , russia .
this work has been supported by the chinese academy of sciences president s international fellowship initiative ( grant no .
2013t2j0011 ) , the japan society for the promotion of science ( grant no.s16019 ) and was initiated through the series of apctp - bltp jinr joint workshops . | we present the new developments in the physics of xyz mesons and glueballs and discuss the connection between properties of the @xmath0 mesons and glueballs , and the structure of the qcd vacuum .
it is shown that the mixing between quark and gluon degree of freedoms induced by the nontrivial topological structure of the qcd vacuum might affect strongly the properties of the exotic hadrons . a new interpretation of the @xmath0 mesons as the mixing between radial excitations of light quark - antiquark systems and states with heavy quark content is suggested .
we introduce also a new mechanism of quarkonia production in hadron - hadron collisions coming from the mixing of light and heavy quark sectors of qcd .
this mechanism is based on the nontrivial topological structure of the qcd vacuum and might explain the polarization puzzle in inclusive @xmath1 production in high energy reactions . the possibility of big changes of the scalar and pseudoscalar glueball masses in the hot gluonic plasma is established and their consequences in the phase structure of qcd analyzed . |
the nonlinearity ( power dependence ) of the microwave ( rf ) surface impedance @xmath4 of the high-@xmath5 superconductors is important for both practical applications and fundamental materials understanding .
nonlinearity in the surface impedance not only degrades the power handling ability of devices but also causes intermodulation and harmonic generation problems at moderate power levels . there has been a large amount of effort to investigate the nonlinear @xmath4 , but the origins of these phenomena are still not understood . since most of the measurements have been performed on photolithographically patterned stripline or coplanar waveguide devices , one of the suggested explanations for the nonlinearity is defects and damage caused by patterning .
all patterned devices have some current crowding near their edges . for example
, stripline and microstrip devices have current distributions that peak sharply near their edges , where the patterning processes have a relatively large influence . a sem picture focusing on the edge of a patterned ybco thin film is shown in fig . [ fig1 ] .
the edge is not straight as can be seen in the picture and the roughness of the edge is of the order of a superconducting penetration depth .
this edge roughness results from the initial imperfections of the photoresist pattern and from nonuniformities of the etching process .
the edge roughness led to the speculation that the condition of the ybco line edge was influencing the power dependence , and this was the initial motivation for this work .
the bright particles on the surface of the ybco film are the well - known copper oxide outgrowths that are present in all films whose stoichiometry is not exact@xcite .
we have found that outgrowths do not adversely affect the microwave properties and therefore it is unnecessary to make large efforts to deposit films with exactly the correct stoichiometry .
we have also observed that improvement of the surface morphology to make smoother films than that shown in fig .
[ fig1 ] does not improve their microwave properties either with regard to their residual @xmath0 or their power dependence .
in the present work , we used both a sapphire dielectric resonator and a stripline resonator to directly compare the power dependence of the microwave surface resistance @xmath0 of the _ same ybco thin films , first unpatterned in the dielectric resonator and then patterned in the stripline resonator .
experimental procedures and results are shown in the following sections . to help understand our experimental results better ,
we have also modeled the @xmath0 as measured in the dielectric resonator and in the stripline resonator using the following simple phenomenological assumption , @xmath6 @xmath7 where @xmath8 and @xmath9 are , respectively , the real part of the complex local resistivity and penetration depth , @xmath10 is the local current density in the film , and @xmath11 and @xmath12 are the corresponding nonlinear coefficients . _
in order to study the effects of defects and damage introduced in the patterning process on the power dependence of the microwave surface resistance of ybco thin films , we measured the nonlinear microwave frequency surface resistance at 10.7ghz of unpatterned , 2-inch - diameter , ybco films using a specially designed sapphire dielectric resonator .
the unpatterned films were then patterned using standard photolithography and wet chemical etching with 0.1% phosphoric acid to make stripline resonators with resonant frequencies @xmath13 = @xmath14@xmath15 , where @xmath16 , ... and @xmath15 = 1.5ghz .
the patterning and dicing procedures are described in fig .
the microwave surface resistance was then measured as a function of rf power for the patterned striplines .
comparing our results `` before '' and `` after '' the patterning process , directly clarifies whether the patterning contributes to the power dependence of the microwave surface resistance observed for ybco thin films .
we have designed a sapphire dielectric resonator shown in fig .
[ fig3 ] to measure the microwave surface resistance and especially the power dependence of unpatterned ybco thin films .
the resonator incorporates some unique features described below that allow high - power measurements to be carried out without heating the sample via dissipated power .
figure [ fig4 ] is a cut - away view of the dielectric resonator .
the resonator was made from a @xmath17-inch - diameter and @xmath18-inch - high cylindrical sapphire puck that is grounded by two 2-inch - diameter ybco wafers ( see left hand side of fig .
[ fig2 ] ) .
the resonator has fixed output coupling and in - situ - adjustable input coupling to ensure critical input coupling ( no reflection at the input end ) for high - power measurements .
the resonator is packaged in an oxygen - free copper case , as shown in fig .
special thermal contacts were designed to avoid heating problems for high - power experiments .
a copper pressure plate was employed on top of the upper ybco wafer .
springs were used ( as shown in fig .
[ fig3 ] ) to ensure that the pressure on the ybco is independent of temperature .
a gold - plated copper foil was soldered onto the copper pressure plate to enhance thermal conduction to the copper case from the top ybco film which was more likely to be affected by heating than the bottom wafer that was anchored to the copper base plate .
thermal - conducting grease was applied between the sapphire puck and the two grounding wafers .
the whole package was operated in a vacuum environment in a cryocooler with the bottom plate bolted to the cold finger with an indium foil for thermal contact .
a temperature sensor and a 25-@xmath19 heater were mounted on the copper case of the resonator with indium foil to monitor and control the temperature .
the sapphire dielectric resonator was operated at the @xmath20 mode for which the center frequency was 10.7ghz .
the loaded quality factor @xmath21 , reflection ( @xmath22 , @xmath23 ) and transmission ( @xmath24 ) coefficients were measured as a function of input power and temperature .
the input coupling loop was tuned in - situ to maintain critical coupling at each temperature and input power level .
critical coupling on the input maximizes the circulating power and rf magnetic field in the resonator .
the output coupling loop was fixed and weakly coupled during each measurement .
the unloaded quality factor @xmath25 was obtained from the measured @xmath26 parameters and @xmath21 , and the surface resistance was deduced from @xmath0 = @xmath27 , where @xmath28 is the calculated geometric factor of the resonator.@xcite the results for @xmath0 are presented as a function of peak magnetic field in the resonator . the @xmath0 measured by this method is the average of @xmath0 in the top and bottom ybco films .
the @xmath21 was measured using a standard frequency - domain 3-db bandwidth method at a low input power where the @xmath29 is independent of power . at higher input power levels ,
a time - domain method was employed to reduce heating effects . in this method , a cw microwave pulse with enough duration to fully charge the resonator was sent into the resonator . after the input pulse is turned off , we can calculate @xmath21 from the decay rate of the output signal that is proportional to @xmath30 $ ] . by use of the time - domain method with this special thermal design , the @xmath0
could be measured without heating problems for input powers up to 40dbm , which corresponds to a rf peak magnetic field of @xmath31100oe .
surface resistance measurements on two pairs of unpatterned ybco films were carried out for temperatures in the range 30 to 80k . in the analysis of the experimental data
, we have , as is usual , treated the losses of the ybco superconducting ground planes as a perturbation that does not affect the field or current distribution , while ignoring the losses due to the sapphire puck itself and to the copper case . the assumption that only the ybco films contribute to the power loss is expected to be valid for the following reasons .
a good - quality sapphire crystal is almost lossless at the cryogenic temperature where we operated , since the @xmath32(loss tangent ) is estimated@xcite to be on the order of @xmath33 . for ybco films ,
a @xmath25 of about @xmath34 is observed at low temperatures .
a pair of 2-inch - diameter nb films have yielded @xmath25 values greater than @xmath35 . that the nb films gave a higher @xmath29 than those of ybco indicates that the loss in the sapphire puck is negligible compared to that of the ybco films .
we estimated that the loss in the copper case is very small and yields a @xmath29 greater than @xmath36 , much higher than that of the ybco films .
after the microwave surface resistance of the unpatterned ybco films was measured with the sapphire dielectric resonator , the films were patterned and stripline resonators were made .
a standard patterning process with wet chemical etching was employed . as shown in fig .
[ fig2 ] , each 2-inch wafer was made into four striplines and four ground planes .
one piece from each of the top and the bottom ybco wafers used with the dielectric resonator in fig .
[ fig3 ] was measured .
a description of the stripline resonator used here can be found in detail elsewhere.@xcite similar to the dielectric resonator method , the loaded quality factor @xmath21 was measured for the stripline as a function of microwave input power up to 35dbm , corresponding to a rf peak magnetic field of 1000oe in the resonator .
one advantage of the stripline resonator is that the @xmath0 at different overtone frequencies can be easily measured . by measuring overtones
, we could measure the @xmath29 at frequencies close to that of the dielectric resonator even though the fundamental resonance of the stripline resonator is 1.5ghz .
the @xmath0 as a function of rf peak magnetic field @xmath37 at different temperatures was measured for two pairs of unpatterned ybco films , denoted as pair 1 and 2 ( see fig . [ fig3 ] ) .
the results are shown in fig .
[ fig5 ] . for pair 1 , no nonlinearity in @xmath0 of the unpatterned wafers
was observed up to 80k even at + 40dbm , the maximum available power , and corresponding to @xmath38 values up to 200oe . as mentioned above
, @xmath0 obtained for the unpatterned wafers is an average surface resistance of the pair of films .
figure [ fig6](a ) compares @xmath0 of the unpatterned and patterned films at 35k for pair 1 .
the open circles represent the @xmath0 of the patterned film from the top wafer measured in the dielectric resonator , and the crosses denote the @xmath0 of the patterned film from the bottom wafer .
the average of those two curves ( dashed line ) agrees with the @xmath0 of the unpatterned wafers very well and suggests that there is no nonlinearity for @xmath37 up to 500oe .
figure [ fig6](b ) shows the results at 75k for pair 1 . for the case of stripline resonators , because of package modes interfering with some of the overtone modes , the @xmath0 could not be measured at the mode closest in frequency to that of the dielectric resonator
. therefore @xmath0 was measured at the closest frequencies possible .
the results for @xmath0 were then scaled assuming @xmath0 @xmath39 @xmath40 . for pair 1 the @xmath0 of the unpatterned films
was measured at 10.7ghz , and the @xmath0 of the two patterned films was measured at 12ghz and 7.5ghz . from the results in fig .
[ fig6 ] , we conclude that for pair 1 , no degradation in the residual surface resistance was observed in the patterned stripline films up to a @xmath37 of @xmath41 200oe .
no power dependence was observed up to the maximum input power available for the unpatterned films , and no effects of patterning on the power handling of ybco films could be found either at 35k or at 75k for pair 1 .
we grew another pair of 2-inch - diameter ybco wafers pair 2 , under a different deposition temperature with the intention to decrease the power handling of the microwave surface resistance .
for pair 2 at 35k no power dependence of the @xmath0 was observed in the dielectric resonator up to the maximum input power , similar to the results obtained for pair 1 .
however , at @xmath42 75k , a nonlinearity in the @xmath0 was observed above @xmath37 @xmath41 100oe .
results for the patterned films are shown together with that of the unpatterned films in fig .
[ fig7 ] . in this case
, @xmath0 was measured at 7.5 ghz for both patterned films in the stripline resonator .
the measured residual surface resistance of the unpatterned and patterned films agrees to within 10@xmath43 .
the @xmath44 for the unpatterned films actually turns up at a somewhat lower @xmath37 than for the patterned films in the stripline resonator .
the apparently better power handling of the stripline resonator can be explained using the same surface impedance by a simple model introduced below .
the @xmath45 is the rf peak magnetic field in the two types of resonators , but the two types of resonators have very different field distributions . in the next section ,
we use the different current / field profiles for both the sapphire dielectric and stripline resonators to model the measured power dependence of the unpatterned and patterned films .
the current profiles in the films for the dielectric and stripline resonators are different , as shown in fig .
[ fig8 ] . for the dielectric resonator with a film diameter much larger than the diameter of the sapphire puck so that the edge effects can be ignored , the current density @xmath46 is in the @xmath47 direction and has a dependence on radius given by@xcite @xmath48 where @xmath49 , @xmath50 , @xmath51 , @xmath52 are various bessel functions , @xmath53 , @xmath54 and @xmath55 are constants determined by the geometry of the resonator , _ a is the radius of the sapphire puck and @xmath56 is the radius of the superconducting films . the current density
thus peaks near the center of the film as shown in fig .
[ fig8 ] . for the stripline resonator
, the current density @xmath46 has been calculated numerically in ref . for the purpose of simplicity in the calculation
, we approximated @xmath57 with the following analytical form,@xcite @xmath58}^{-\frac{1}{2 } } & x\ll w\\ { j_s(0)}(\frac{1.165}{\lambda_0})(wb)^{\frac{1}{2}}\exp(-\frac{(w/2-|x|)b}{{\lambda_0}^2 } ) & x\approx w \end{array } \right .
\label{eq1.4}\ ] ] where , @xmath59 and @xmath60 are , respectively , the width and thickness of the film .
this analytical approximation agrees with the numerically calculated current distribution to within 5@xmath43 in our case .
thus @xmath57 peaks sharply at the edges of the patterned film for the stripline resonator . for the dielectric resonator ,
as is obvious in eq .
( [ eq1.3 ] ) , @xmath57 is related directly to the rf peak magnetic field @xmath37 . for the stripline resonator
, @xmath37 can be calculated numerically from the current distribution in the stripline.@xcite for the same maximum field @xmath37 in the dielectric and stripline resonators , the portion of the film carrying a high current density is larger for the dielectric resonator due to the broader peak in the @xmath61 distribution of the dielectric resonator .
therefore an apparently better power handling in the stripline resonator is not surprising . _
the quality factors @xmath25(@xmath37 ) and in turn @xmath0(@xmath37 ) of both resonators have been modeled as a function of rf peak magnetic field using the current distributions given above , with the assumption for the impedance is given by eqs .
( [ eq1.1 ] ) and ( [ eq1.2 ] ) , and @xmath62 where @xmath63 is the total energy stored and @xmath64 is the power dissipated in the resonator .
the parameters @xmath65 and @xmath66 were taken from experimentally measured values , @xmath67@xmath19 m , @xmath66 = 0.2@xmath68 m , that fit the low - field , linear part of @xmath0 and the temperature dependence of the resonant frequency .
the nonlinear parameters of resistivity and penetration depth used in the calculations in the superconducting phase were @xmath69@xmath19@xmath70 and @xmath71@xmath70 .
these parameters were taken to fit the @xmath0 obtained from the dielectric resonator measurements . the unloaded quality factor for the stripline resonator
was then calculated with the same set of parameters . at low rf field ,
calculated values of @xmath25 of @xmath72 and 6000 , respectively , were obtained for the dielectric and stripline resonators , and these values are consistent with the measured values .
the modeled results also show that the power dependence of @xmath0 for both the dielectric and stripline resonators is quantitatively consistent with the measured power dependence for both resonators .
the modeled @xmath0 of the dielectric resonator turns up at a smaller value of @xmath45 than that of the stripline resonator , which agrees with the experimental result well as shown in fig .
we have measured the microwave surface resistance of the _ same ybco thin films , before and after patterning , using a sapphire dielectric resonator and a stripline resonator .
a phenomenological model of the @xmath0@xmath73 in the two cases is presented .
based on this model , the calculated results fit the experimental data well in both cases using the same materials parameters .
therefore , we conclude that our results are consistent with no damage or degradation of the power dependence of the microwave surface resistance due to the patterning of the films .
the procedure described in this paper can also be used to test other patterning processes such as ion beam etching , etc . _
this work was supported by the air force office of scientific research .
the authors express their gratitude to bob koneizcka and earle macedo for applying their tireless efforts and effective skills to the preparation of the devices used in this study , and to peter murphy for providing the sem pictures used in this study . | the effect of the patterning process on the nonlinearity of the microwave surface resistance @xmath0 of ybco thin films is investigated . with the use of a sapphire dielectric resonator and a stripline resonator , the microwave @xmath0 of ybco thin films
was measured before and after the patterning process , as a function of temperature and the rf peak magnetic field in the film .
the microwave loss was also modeled , assuming a @xmath1 dependence of @xmath2 on current density @xmath3 .
experimental and modeled results show that the patterning has no observable effect on the microwave residual @xmath0 or on the power dependence of @xmath0 .
= 10000 |
the problem of how to construct a bivariate random vector @xmath2 with log - normal marginals @xmath3 , @xmath4 and correlation coefficient @xmath5 is well known in the history of dependence modeling , partially because of its relevance to risk management practice .
the short answer is : there is no such model ; see embrechts et al .
@xcite who studied these kinds of problems in terms of copulas . problems of this kind were brought to risklab at eth zurich by the insurance industry in the mid-1990s when dependence was thought of in terms of correlation ( matrices ) . for further background on quantitative risk management ,
see mcneil et al .
now , almost 20 years later , copulas are a well established tool to quantify dependence in multivariate data and to construct new multivariate distributions .
their use has become standard within industry and regulation .
nevertheless , dependence is still summarized in terms of numbers [ as opposed to ( copula ) functions ] , so - called _ measures of association_. although there are various ways to compute such numbers in dimension @xmath6 , measures of association are still most widely used in the bivariate case @xmath7 .
a popular measure of association is tail dependence .
it is important for applications in quantitative risk management as it measures the strength of dependence in either the lower - left or upper - right tail of the bivariate distribution , the regions quantitative risk management is mainly concerned with .
we were recently asked the following question which is in the same spirit as the log - normal correlation problem if one replaces `` correlation '' by `` tail dependence '' ; see section [ sec : lambda:1 ] for a definition .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for which @xmath8 $ ] is the matrix _
@xmath9 _ a matrix of pairwise ( either lower or upper ) tail - dependence coefficients ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ intrigued by this question , we more generally consider the following _ tail - dependence compatibility problem _ in this paper : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ when is a given matrix in @xmath10^{d\times d}$ ] the matrix of pairwise ( either lower or upper ) tail - dependence coefficients ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in what follows , we call a matrix of pairwise tail - dependence coefficients a _ tail - dependence matrix_. the compatibility problems of tail - dependence coefficients were studied in @xcite . in particular , when @xmath11 , inequalities for the bivariate tail - dependence coefficients have been established ; see joe @xcite , theorem 3.14 , as well as joe @xcite , theorem 8.20 .
the sharpness of these inequalities is obtained in @xcite .
it is generally open to characterize the tail - dependence matrix compatibility for @xmath12 .
our aim in this paper is to give a full answer to the tail - dependence compatibility problem ; see section [ sec : lambda ] . to this end
, we introduce and study _ bernoulli - compatible matrices _ in section [ sec : bern ] .
as a main result , we show that a matrix with diagonal entries being 1 is a compatible tail - dependence matrix if and only if it is a bernoulli - compatible matrix multiplied by a constant . in section [ sec : model ]
, we provide probabilistic models for a large class of tail - dependence matrices , including commonly used matrices in statistics .
section [ sec : con ] concludes . throughout this paper ,
@xmath1 and @xmath13 are positive integers , and we consider an atomless probability space @xmath14 on which all random variables and random vectors are defined .
vectors are considered as column vectors . for two matrices @xmath15 , @xmath16 and @xmath17
are understood as component - wise inequalities .
we let @xmath18 denote the hadamard product , that is , the element - wise product of two matrices @xmath19 and @xmath20 of the same dimension .
the @xmath0 identity matrix is denoted by @xmath21 . for a square matrix @xmath19 ,
@xmath22 represents a diagonal matrix with diagonal entries equal to those of @xmath19 , and @xmath23 is the transpose of @xmath19 .
we denote @xmath24 the indicator function of an event ( random or deterministic ) @xmath25 . @xmath26 and @xmath27 are vectors with all components being 0 and 1 , respectively , as long as the dimension of the vectors is clear from the context .
in this section , we introduce and study the _ bernoulli - compatibility problem_. the results obtained in this section are the basis for the _ tail - dependence compatibility problem _ treated in section [ sec : lambda ] ; many of them are of independent interest , for example , for the simulation of sequences of bernoulli random variables .
a _ bernoulli vector _ is a random vector @xmath28 supported by @xmath29 for some @xmath30 .
the set of all @xmath1-bernoulli vectors is denoted by @xmath31 .
equivalently , @xmath32 is a bernoulli vector if and only if @xmath33 for some @xmath34 $ ] , @xmath35 .
note that here we do not make any assumption about the dependence structure among the components of @xmath28 .
bernoulli vectors play an important role in credit risk analysis ; see , for example , bluhm and overbeck @xcite and bluhm et al .
@xcite , section 2.1 . in this section
, we investigate the following question which we refer to as the _ bernoulli - compatibility problem_. [ q : b : compatibility ] given a matrix @xmath36^{d\times d}$ ] , can we find a bernoulli vector @xmath28 such that @xmath37 $ ] ? for studying the bernoulli - compatibility problem , we introduce the notion of bernoulli - compatible matrices .
a @xmath0 matrix @xmath20 is a _ bernoulli - compatible matrix _ , if @xmath37 $ ] for some @xmath38 .
the set of all @xmath0 bernoulli - compatible matrices is denoted by @xmath39 .
concerning covariance matrices , there is extensive research on the compatibility of covariance matrices of bernoulli vectors in the realm of statistical simulation and time series analysis ; see , for example , chaganty and joe @xcite .
it is known that , when @xmath40 , the set of all compatible @xmath1-bernoulli correlation matrices is strictly contained in the set of all correlation matrices .
note that @xmath41={\operatorname{cov}}(\mathbf{x})+{\mathbb{e}}[\mathbf { x}]{\mathbb{e}}[\mathbf{x}]^\top$ ] .
hence , question [ q : b : compatibility ] is closely related to the characterization of compatible bernoulli covariance matrices .
before we characterize the set @xmath39 in section [ chara : bcm ] , and thus address question [ q : b : compatibility ] , we first collect some facts about elements of @xmath39 . [
prop : prelim ] let @xmath42 .
then : @xmath36^{d\times d}$ ] .
@xmath43 for @xmath44 and @xmath45 .
@xmath46 for @xmath47 $ ] , that is , @xmath39 is a convex set .
@xmath48 , that is , @xmath39 is closed under the hadamard product .
@xmath49 and @xmath50 . for any @xmath51^d$
] , the matrix @xmath52 where @xmath53 for @xmath54 and @xmath55 , @xmath44 . write @xmath56 $ ] and @xmath57 $ ] for @xmath58 , and @xmath28 and @xmath59 are independent .
this directly follows from the frchet
hoeffding bounds ; see mcneil et al .
@xcite , remark 7.9 .
let @xmath60 be a bernoulli random variable independent of @xmath61 , and let @xmath62 .
then @xmath63 , and @xmath64=t { \mathbb{e}}[\mathbf{x}\mathbf{x}^\top]+(1-t){\mathbb{e}}[\mathbf{y}\mathbf{y}^\top ] = t b_1+(1-t)b_2 $ ] .
hence , @xmath65 .
let @xmath66 , @xmath67 .
then @xmath68 let @xmath69 .
it follows that @xmath63 and @xmath64={\mathbb{e}}[(\mathbf{x}\circ\mathbf{y } ) ( \mathbf{x}\circ\mathbf{y})^\top ] = { \mathbb{e}}[(\mathbf{x}\mathbf{x}^\top)\circ(\mathbf{y}\mathbf{y}^\top ) ] = { \mathbb{e}}[\mathbf{x}\mathbf{x}^\top]\circ{\mathbb{e}}[\mathbf{y}\mathbf { y}^\top]=b_1\circ b_2 $ ] .
hence , @xmath48 .
consider @xmath70 .
then @xmath71\in{\mathcal{b}}_d$ ] and similarly for @xmath72 .
consider @xmath38 with independent components and @xmath73={\mathbf}p$ ] .
we are now able to give a characterization of the set @xmath39 of bernoulli - compatible matrices and thus address question [ q : b : compatibility ] .
[ thm : bern ] @xmath39 has the following characterization : @xmath74 that is , @xmath39 is the convex hull of @xmath75 . in particular , @xmath39 is closed under convergence in the euclidean norm .
denote the right - hand side of ( [ bern-1 ] ) by @xmath76 . for @xmath77 , write @xmath37 $ ] for some @xmath78 .
it follows that @xmath79 hence @xmath80 .
. then @xmath38 and @xmath82={\mathbf}p { \mathbf}p^\top\in{\mathcal{b}}_d$ ] .
by proposition [ prop : prelim ] , @xmath39 is a convex set which contains @xmath83 , hence @xmath84 . in summary , @xmath85 . from ( [ bern-1 ] ) , we can see that @xmath39 is closed under convergence in the euclidean norm .
a matrix @xmath20 is _ completely positive _ if @xmath86 for some ( not necessarily square ) matrix @xmath87 .
denote by @xmath88 the set of completely positive matrices .
it is known that @xmath88 is the convex cone with extreme directions @xmath89^d\}$ ] ; see , for example , rschendorf @xcite and berman and shaked - monderer @xcite .
we thus obtain the following result .
[ cpm ] any bernoulli - compatible matrix is completely positive .
[ rem : bd ] one may wonder whether @xmath37 $ ] is sufficient to determine the distribution of @xmath28 , that is , whether the decomposition @xmath90 is unique for distinct vectors @xmath91 in @xmath29 . while the decomposition is trivially unique for @xmath7 ,
this is in general false for @xmath40 , since there are @xmath92 parameters in ( [ op1 ] ) and only @xmath93 parameters in @xmath20 .
the following is an example for @xmath11 .
let @xmath94 thus , by combining the above two decompositions , @xmath95 has infinitely many different decompositions of the form ( [ op1 ] ) . note that , as in the case of completely positive matrices , it is generally difficult to find decompositions of form ( [ op1 ] ) for a given matrix @xmath20 . in this section , we study the convex cone generated by @xmath39 , denoted by @xmath96 : @xmath97 the following proposition is implied by proposition [ prop : prelim ] and theorem [ thm : bern ] .
[ prop : bern ] @xmath96 is the convex cone with extreme directions @xmath98 .
moreover , @xmath96 is a commutative semiring equipped with addition @xmath99 and multiplication @xmath100 .
it is obvious that @xmath101 .
one may wonder whether @xmath96 is identical to @xmath88 , the set of completely positive matrices .
as the following example shows , this is false in general for @xmath102 .
note that @xmath103 also satisfies proposition [ prop : prelim ] , part ( ii ) .
now consider @xmath104 with @xmath105 for some @xmath106 .
clearly , @xmath107 , but @xmath108 contradicts proposition [ prop : prelim ] , part ( ii ) , hence @xmath109 . for the following result ,
we need the notion of diagonally dominant matrices .
a matrix @xmath110 is called _ diagonally dominant _ if , for all @xmath35 , @xmath111 .
[ ddm ] let @xmath112 be the set of nonnegative , diagonally dominant @xmath113-matrices .
then @xmath114 . for @xmath44 ,
let @xmath115 where @xmath116 .
it is straightforward to verify that the @xmath117- , @xmath118- , @xmath119- and @xmath120-entries of the matrix @xmath121 are 1 , and the other entries are 0 . for @xmath122 ,
let @xmath123 by proposition [ prop : bern ] , @xmath124 .
it follows that @xmath125 for @xmath54 and @xmath126 .
therefore , @xmath127 , which , by proposition [ prop : bern ] , is in @xmath96 . for studying the tail - dependence compatibility problem in section [ sec : lambda ] ,
the subset @xmath128 of @xmath96 is of interest .
it is straightforward to see from proposition [ prop : prelim ] and theorem [ thm : bern ] that @xmath129 is a convex set , closed under the hadamard product and convergence in the euclidean norm .
these properties of @xmath129 will be used later .
the notion of tail dependence captures ( extreme ) dependence in the lower - left or upper - right tails of a bivariate distribution . in what follows ,
we focus on lower - left tails ; the problem for upper - right tails follows by a reflection around @xmath130 , that is , studying the survival copula of the underlying copula .
the _ ( lower ) tail - dependence coefficient _ of two continuous random variables @xmath131 and @xmath132 is defined by @xmath133 given that the limit exists .
if we denote the copula of @xmath2 by @xmath134 , then @xmath135 clearly , @xmath136 $ ] , and @xmath137 only depends on the copula of @xmath2 , not the marginal distributions . for
virtually all copula models used in practice , the limit in ( [ def : tail : dep ] ) exists ; for how to construct an example where @xmath137 does not exist ; see kortschak and albrecher @xcite .
let @xmath32 be a random vector with continuous marginal distributions .
the _ tail - dependence matrix _ of @xmath28 is @xmath138 , where @xmath139 is the tail - dependence coefficient of @xmath140 and @xmath141 , @xmath44 .
we denote by @xmath142 the set of all tail - dependence matrices .
the following proposition summarizes basic properties of tail - dependence matrices .
its proof is very similar to that of proposition [ prop : prelim ] and is omitted here . for any @xmath143
, we have that : @xmath144 .
@xmath145 for @xmath47 $ ] , that is , @xmath142 is a convex set .
@xmath146 with @xmath147 and @xmath148 .
as we will show next , @xmath142 is also closed under the hadamard product .
[ prop : hadamard : prod ] let @xmath149 and @xmath150
. then @xmath151 .
note that it would be sufficient to show the result for @xmath152 , but we provide a general construction for any @xmath153 . for each @xmath154 , let @xmath155 be a @xmath1-dimensional copula with tail - dependence matrix @xmath156 .
furthermore , let @xmath157 , @xmath158 $ ] .
it follows from liebscher @xcite that @xmath159 is a copula ; note that @xmath160 for independent random vectors @xmath161 , @xmath154 .
the @xmath118-entry @xmath139 of @xmath162 corresponding to @xmath134 is thus given by @xmath163 where @xmath164 denotes the @xmath118-margin of @xmath155 and @xmath165 denotes the @xmath118th entry of @xmath156 , @xmath154 . in this section , we investigate the following question .
[ q : characterization ] given a @xmath0 matrix @xmath166^{d\times d}$ ] , is it a tail - dependence matrix ?
the following theorem fully characterizes tail - dependence matrices , and thus provides a theoretical ( but not necessarily practical ) answer to question [ q : characterization ] .
[ thm : main : characterization ] a square matrix with diagonal entries being 1 is a tail - dependence matrix if and only if it is a bernoulli - compatible matrix multiplied by a constant .
equivalently , @xmath167 .
we first show that @xmath168 . for each @xmath169 , suppose that @xmath134 is a copula with tail - dependence matrix @xmath162 and @xmath170 .
let @xmath171 . by definition , @xmath172\ ] ] and
@xmath173.\ ] ] since @xmath129 is closed and @xmath174/u\in{\mathcal{b}}^i_d$ ] , we have that @xmath175
. now consider @xmath176 . by definition of @xmath129
, each @xmath177 can be written as @xmath37/p$ ] for an @xmath38 and @xmath73=(p,\ldots , p)\in(0,1]^d$ ] .
let @xmath178 $ ] , @xmath179 be independent and @xmath180 we can verify that for @xmath47 $ ] and @xmath35 , @xmath181 that is , @xmath182 are @xmath183$]-distributed .
let @xmath139 be the tail - dependence coefficient of @xmath184 and @xmath185 , @xmath44 .
for @xmath44 we obtain that @xmath186.\end{aligned}\ ] ] as a consequence , the tail - dependence matrix of @xmath187 is @xmath20 and @xmath188 .
it follows from theorem [ thm : main : characterization ] and proposition [ prop : bern ] that @xmath142 is the `` 1-diagonals '' cross - section of the convex cone with extreme directions @xmath83 .
furthermore , the proof of theorem [ thm : main : characterization ] is constructive .
as we saw , for any @xmath189 , @xmath59 defined by ( [ smodel ] ) has tail - dependence matrix @xmath20 .
this interesting construction will be applied in section [ sec : model ] where we show that commonly applied matrices in statistics are tail - dependence matrices and where we derive the copula of @xmath59 . from the fact that @xmath190 and @xmath191 is closed under the hadamard product
[ see proposition [ prop : prelim ] , part ( iv ) ] , proposition [ prop : hadamard : prod ] directly follows .
note , however , that our proof of proposition [ prop : hadamard : prod ] is constructive . given tail - dependence matrices and corresponding copulas
, we can construct a copula @xmath134 which has the hadamard product of the tail - dependence matrices as corresponding tail - dependence matrix .
if sampling of all involved copulas is feasible , we can sample @xmath134 ; see figure [ fig : liebscher ] for examples
. 0.999 - 13 ) by calling ` demo(tail_compatibility ) ` . ] ) for @xmath192 being a clayton copula with parameter @xmath193 ( @xmath194 ) and @xmath195 being a @xmath196 copula with parameter @xmath197 [ tail - dependence coefficient @xmath198 .
by proposition [ prop : hadamard : prod ] , the tail - dependence coefficient of ( [ eq : liebscher : cop ] ) is thus @xmath199 .
right - hand side : @xmath192 as before , but @xmath195 is a survival marshall olkin copula with parameters @xmath200 , so that @xmath201 . ]
theorem [ thm : main : characterization ] combined with corollary [ cpm ] directly leads to the following result .
every tail - dependence matrix is completely positive , and hence positive semi - definite .
furthermore , theorem [ thm : main : characterization ] and proposition [ ddm ] imply the following result .
[ coro : ddm ] every diagonally dominant matrix with nonnegative entries and diagonal entries being @xmath202 is a tail - dependence matrix .
note that this result already yields the if - part of proposition [ prop : example ] below .
we now consider the following three types of matrices @xmath138 which are frequently applied in multivariate statistics and time series analysis and show that they are tail - dependence matrices .
equicorrelation matrix with parameter @xmath203 $ ] : @xmath204 , @xmath205 .
ar(1 ) matrix with parameter @xmath8 $ ] : @xmath206 , @xmath205 .
ma(1 ) matrix with parameter @xmath207 $ ] : @xmath208 , @xmath205 .
chaganty and joe @xcite considered the compatibility of correlation matrices of bernoulli vectors for the above three types of matrices and obtained necessary and sufficient conditions for the existence of compatible models for @xmath11 . for the tail - dependence compatibility problem that we consider in this paper , the above three types of matrices
are all compatible , and we are able to construct corresponding models for each case .
let @xmath162 be the tail - dependence matrix of the @xmath1-dimensional random vector @xmath209 where @xmath210 $ ] , @xmath38 and @xmath179 are independent . for @xmath8 $ ] ,
if @xmath28 has independent components and @xmath211=\cdots={\mathbb{e}}[x_d]=\alpha$ ] , then @xmath162 is an equicorrelation matrix with parameter @xmath212 ; that is , is a tail - dependence matrix . for @xmath8 $ ] , if @xmath213 , @xmath35 , for independent @xmath214 random variables @xmath215 , then @xmath162 is an ar(1 ) matrix with parameter @xmath212 ; that is , is a tail - dependence matrix . for @xmath216 $ ] , if @xmath217\}}$ ] , @xmath35 , for @xmath218 $ ] , then @xmath162 is an ma(1 ) matrix with parameter @xmath212 ; that is , is a tail - dependence matrix .
we have seen in the proof of theorem [ thm : main : characterization ] that if @xmath211=\cdots={\mathbb{e}}[x_d]=p$ ] , then @xmath59 defined through ( [ smodel2 ] ) has tail - dependence matrix @xmath41/p$ ] .
write @xmath219 and note that @xmath220 , @xmath35 , is always guaranteed . for @xmath106
, we have that @xmath221=\alpha^2 $ ] and thus @xmath222 .
this shows that @xmath162 is an equicorrelation matrix with parameter @xmath212 . for @xmath223 ,
we have that @xmath224&=&{\mathbb{e}}\biggl [ \prod _ { k = i}^{i+d-1 } z_k \prod _ { l = j}^{j+d-1 } z_l \biggr]={\mathbb{e}}\biggl [ \prod _ { k = i}^{j-1 } z_k \biggr]{\mathbb{e}}\biggl [ \prod_{k = j}^{i+d-1 } z_k \biggr]{\mathbb{e}}\biggl [ \prod_{k = i+d}^{j+d-1 } z_k \biggr ] \\ & = & \alpha^{j - i } \alpha^{i+d - j}\alpha^{j - i}= \alpha^{j - i+d}\end{aligned}\ ] ] and @xmath225={\mathbb{e}}[x_i^2]=\alpha^d$ ] . hence , @xmath226 for @xmath227 . by symmetry ,
@xmath206 for @xmath54 .
thus , @xmath162 is an ar(1 ) matrix with parameter @xmath212 .
for @xmath223 , note that @xmath228 , so @xmath224&=&{\mathbb{p}}\bigl(z\in\bigl[(j-1 ) ( 1-\alpha ) , ( i-1 ) ( 1-\alpha)+1\bigr]\bigr ) \\ & = & { \mathrm{i}}_{\{j = i+1\ } } { \mathbb{p}}\bigl(z\in\bigl[i(1-\alpha),(i-1 ) ( 1-\alpha)+1\bigr ] \bigr)={\mathrm{i}}_{\ { j = i+1\}}\frac{\alpha } d\end{aligned}\ ] ] and @xmath225={\mathbb{e}}[x_i^2]=\frac{1}d$ ] .
hence , @xmath229 for @xmath223 . by symmetry ,
@xmath230 for @xmath54 .
thus , @xmath162 is an ma(1 ) matrix with parameter @xmath212 .
theorem [ thm : main : characterization ] gives a characterization of tail - dependence matrices using bernoulli - compatible matrices and ( [ smodel ] ) provides a compatible model @xmath59 for any tail - dependence matrix @xmath231/p)$ ] .
it is generally not easy to check whether a given matrix is a bernoulli - compatible matrix or a tail - dependence matrix ; see also remark [ rem : bd ] .
therefore , we now study the following question . [ q : model : for : y ] how can we construct a broader class of models with flexible dependence structures and desired tail - dependence matrices ? to enrich our models , we bring random matrices with bernoulli entries into play . for @xmath232 ,
let @xmath233 that is , @xmath234 is the set of @xmath235 random matrices supported in @xmath236 with each row being _ mutually exclusive _ ; see dhaene and denuit @xcite .
furthermore , we introduce a transformation @xmath237 on the set of square matrices , such that , for any @xmath44 , the @xmath118th element @xmath238 of @xmath239 is given by @xmath240 that is , @xmath237 adjusts the diagonal entries of a matrix to be 1 , and preserves all the other entries . for a set @xmath241 of square matrices , we set @xmath242
. we can now address question [ q : model : for : y ] .
[ thm : main : tdm : y ] let @xmath243 for an @xmath13-dimensional copula @xmath244 with tail - dependence matrix @xmath162 and let @xmath245 for a @xmath1-dimensional copula @xmath246 with tail - dependence matrix @xmath21 .
furthermore , let @xmath247 such that @xmath248 are independent and let @xmath249 where @xmath250 with @xmath251 , @xmath35
. then @xmath59 has tail - dependence matrix @xmath252)$ ] .
write @xmath253 , @xmath254 , @xmath255 , @xmath138 and @xmath256 . then , for all @xmath35 , @xmath257 clearly , @xmath59 has @xmath183 $ ] margins .
we now calculate the tail - dependence matrix @xmath258 of @xmath259 for @xmath106 . by our independence assumptions , we can derive the following results : @xmath260 , where @xmath261 denotes the @xmath118th margin of @xmath246 .
as @xmath262 has tail - dependence matrix @xmath21 , we obtain that @xmath263 @xmath264 , and thus @xmath265 similarly , we obtain that @xmath266 @xmath267 $ ] so that @xmath268\\ & = & { \mathbb{e}}\biggl [ \sum _ { k=1}^m\sum_{l=1}^mx_{ik } \lambda_{kl}x_{jl } \biggr ] \\ & = & \bigl({\mathbb{e}}\bigl[x\lambda x^\top\bigr ] \bigr)_{ij}.\end{aligned}\ ] ] by the law of total probability , we thus obtain that @xmath269 \bigr)_{ij}.\end{aligned}\ ] ] this shows that @xmath270 $ ] and @xmath271 agree on the off - diagonal entries .
since @xmath272 implies that @xmath273 , we conclude that @xmath274)=\gamma$ ] .
a special case of theorem [ thm : main : tdm : y ] reveals an essential difference between the transition rules of a tail - dependence matrix and a covariance matrix .
suppose that for @xmath275 , @xmath276 $ ] is a stochastic matrix ( each row sums to 1 ) , and @xmath243 for an @xmath13-dimensional copula @xmath244 with tail - dependence matrix @xmath138 .
now we have that @xmath277 , @xmath35 in ( [ eq : stoch : rep ] ) . by theorem [ thm : main : tdm : y ] , the tail dependence matrix of @xmath278 is given by @xmath274)$ ] .
one can check the diagonal terms of the matrix @xmath279 by @xmath280 hence , the tail - dependence matrix of @xmath59 is indeed @xmath270 $ ] .
[ transition ] in summary : if an @xmath13-vector @xmath281 has covariance matrix @xmath282 , then @xmath283 has covariance matrix @xmath284 $ ] for any @xmath235 random matrix @xmath285 independent of @xmath281 . if an @xmath13-vector @xmath281 has uniform @xmath10 $ ] margins and tail - dependence matrix @xmath162 , then @xmath283 has tail - dependence matrix @xmath270 $ ] for any @xmath275 independent of @xmath281 such that each row of @xmath285 sums to 1 .
it is noted that the transition property of tail - dependence matrices is more restricted than that of covariance matrices .
the following two propositions consider selected special cases of this construction which are more straightforward to apply .
[ coro - b ] for any @xmath286 and any @xmath287 we have that @xmath288 .
in particular , @xmath289 , and hence @xmath290 . write @xmath291 $ ] for some @xmath292 and consider @xmath293 . as in the proof of theorem [ thm : main
: tdm : y ] ( and with the same notation ) , it follows that for @xmath294 , @xmath295={\mathbb{e}}[w_iw_j\lambda _ { ij}]$ ] .
this shows that @xmath270={\mathbb{e}}[\mathbf{w}\mathbf{w}^\top\circ \lambda]$ ] and @xmath296 agree on off - diagonal entries .
thus , @xmath297 . by taking @xmath298
, we obtain @xmath299 .
the following proposition states a relationship between substochastic matrices and tail - dependence matrices . to this end , let @xmath300 that is , @xmath301 is the set of @xmath235 _ ( row ) substochastic matrices _ ; note that the expectation of a random matrix in @xmath302 is a substochastic matrix .
for any @xmath303 and any @xmath304 , we have that @xmath305 .
in particular , @xmath306 for all @xmath307 and @xmath308 for all @xmath309^d$ ] .
write @xmath310 and let @xmath311 for independent @xmath312 $ ] , @xmath35 , @xmath313 .
it is straightforward to see that @xmath276=q$ ] , @xmath275 with independent rows , and @xmath314 for @xmath35 , so @xmath315 . as in the proof of theorem
[ thm : main : tdm : y ] ( and with the same notation ) , it follows that for @xmath54 , @xmath316 { \mathbb{e}}[x_{jl}]\lambda_{kl}= \sum_{l=1}^m \sum_{k=1}^m q_{ik}q_{jl } \lambda_{kl}.\ ] ] this shows that @xmath317 and @xmath271 agree on off - diagonal entries , so @xmath318 . by taking @xmath319
, we obtain @xmath320 . by taking @xmath321
, we obtain @xmath308 . in this section
, we derive the copulas of ( [ smodel ] ) and ( [ eq : stoch : rep ] ) which are able to produce tail - dependence matrices @xmath41/p$ ] and @xmath274)$ ] as stated in theorems [ thm : main : characterization ] and [ thm : main : tdm : y ] , respectively .
we first address the former .
let @xmath38 , @xmath73=(p,\ldots , p)\in(0,1]^d$ ] .
furthermore , let @xmath210 $ ] , @xmath179 be independent and @xmath322 then the copula @xmath134 of @xmath59 at @xmath323 is given by @xmath324 with the convention @xmath325 . by the law of total probability and our independence assumptions , @xmath326 & = & \sum_{\mathbf{i}\in\{0,1\}^d}{\mathbb{p}}\bigl(pu\le\min_{r : i_r=1 } \{u_r\ }
, p+(1-p)v\le\min_{r : i_r=0}\{u_r\ } , \mathbf{x}=\mathbf{i}\bigr ) \\[-2pt ] & = & \sum_{\mathbf{i}\in\{0,1\}^d}{\mathbb{p}}\biggl(u\le\frac{\min_{r : i_r=1}\{u_r\}}{p } \biggr ) { \mathbb{p}}\biggl(v\le\frac{\min_{r : i_r=0}\ { u_r\}-p}{1-p } \biggr ) { \mathbb{p}}(\mathbf{x}=\mathbf{i});\end{aligned}\ ] ] the claim follows from the fact that @xmath210 $ ] . for deriving the copula of ( [ eq : stoch : rep ] ) , we need to introduce some notation
; see also example [ ex : special : cases ] below . in the following theorem ,
let @xmath327 denote the support of @xmath285 .
for a vector @xmath328^d$ ] and a matrix @xmath329 , denote by @xmath330 the sum of the @xmath331th row of @xmath19 , @xmath35 , and let @xmath332 , and @xmath333 , where @xmath325 .
suppose that the setup of theorem [ thm : main : tdm : y ] holds .
then the copula @xmath134 of @xmath59 in ( [ eq : stoch : rep ] ) is given by @xmath334 by the law of total probability , it suffices to verify that @xmath335 .
this can be seen from @xmath336 & & \qquad={\mathbb{p}}\biggl ( \sum_{k=1}^m a_{jk}u_k+(1-a_j)v_j\le u_j , j=1,\ldots , d \biggr ) \\[-2pt ] & & \qquad={\mathbb{p}}(u_k { \mathrm{i}}_{\{a_{jk}=1\ } } \le u_j , v_j { \mathrm{i}}_{\{a_j=0\}}\le u_j , j=1,\ldots , d,~k=1,\ldots , m ) \\[-2pt ] & & \qquad={\mathbb{p}}\bigl(u_k \le\min_{r : a_{rk}=1}\{u_r\ } , v_j\le u_j { \mathrm{i}}_{\{a_j=0\}}+{\mathrm{i}}_{\{a_j=1\}},\\[-2pt ] & & \qquad\quad j=1,\ldots , d , k=1,\ldots , m \bigr ) \\[-2pt ] & & \qquad={\mathbb{p}}\bigl(u_k \le\min_{r : a_{rk}=1}\{u_r \ } , k=1,\ldots , m \bigr ) { \mathbb{p}}(v_j\le u_j { \mathrm{i}}_{\{a_j=0\}}+{\mathrm{i}}_{\{a_j=1\}},\\[-2pt ] & & \qquad\quad j=1,\ldots , d ) \\[-2pt ] & & \qquad = c^{\mathbf{u}}\bigl(\mathbf{u}_a^*\bigr ) c^{\mathbf{v } } ( { \mathbf{u}}_a).\ ] ] as long as @xmath246 has tail - dependence matrix @xmath21 , the tail - dependence matrix of @xmath59 is not affected by the choice of @xmath246 .
this theoretically provides more flexibility in choosing the body of the distribution of @xmath59 while attaining a specific tail - dependence matrix .
note , however , that this also depends on the choice of @xmath285 ; see the following example where we address special cases which allow for more insight into the rather abstract construction ( [ eq : cop : y ] ) .
[ ex : special : cases ] 1 . for @xmath321 , the copula @xmath134 in ( [ eq : cop : y ] )
is given by @xmath337 note that @xmath338 in equation ( [ eq : cop : y ] ) are indeed vectors in this case .
for @xmath7 , we obtain @xmath339 and therefore a mixture of the frchet hoeffding upper bound @xmath340 @xmath341 , the copula @xmath246 and the independence copula @xmath342 .
if @xmath343 then @xmath134 is simply a mixture of @xmath344 and @xmath345 and does not depend on @xmath346 anymore .
now consider the special case of ( [ eq : case : m1 ] ) where @xmath346 follows the @xmath1-dimensional independence copula @xmath347 and @xmath348 is such that at most one of @xmath349 is 1 [ each randomly with probability @xmath350 and all are simultaneously 0 with probability @xmath351 . then , for all @xmath352^d$ ] , @xmath134 is given by @xmath353 this copula is a conditionally independent multivariate frchet copula studied in yang et al .
. this example will be revisited in section [ sec : ex : qrm ] ; see also the left - hand side of figure [ fig : gamma ] below . for @xmath354 and @xmath281 following a bivariate ( @xmath355 ) @xmath196 copula with kendall s tau equal to 0.75 ( top row ) or a survival marshall olkin copula with parameters @xmath356 ( bottom row ) . for the plots on the left - hand side ,
the number of rows of @xmath285 with one 1 are randomly chosen among @xmath357 , the corresponding rows and columns are then randomly selected among @xmath358 and @xmath359 , respectively .
for the plots on the right - hand side , @xmath285 is drawn from a multinomial distribution with probabilities 0.5 and 0.5 such that each row contains precisely one 1 . ]
\2 . for @xmath355 , @xmath7
, we obtain @xmath360 \\[-8pt ] \nonumber & & { } + c^{\mathbf{v}}(u_1,u_2){\mathbb{p}}\biggl(x={{\left(\matrix { 0 & 0 \cr 0 & 0 } \right)}}\biggr)\\ & & { } + \pi ( u_1,u_2){\mathbb{p}}\biggl(x={{\left(\matrix { 0 & 0 \cr 1 & 0 } \right ) } } \mbox { or } { { \left(\matrix { 0 & 0 \cr 0 & 1 } \right ) } } \mbox { or } { { \left(\matrix { 1 & 0 \cr 0 & 0 } \right ) } } \mbox { or } { { \left(\matrix { 0 & 1 \cr 0 & 0 } \right)}}\biggr).\hspace*{-25pt}\nonumber\end{aligned}\ ] ] figure [ fig : y : m2 ] shows samples of size 2000 from ( [ ex : c : d2:m2 ] ) for @xmath361 and two different choices of @xmath281 ( in different rows ) and @xmath285 ( in different columns ) . from theorem [ thm :
main : tdm : y ] , we obtain that the off - diagonal entry @xmath362 of the tail - dependence matrix @xmath363 of @xmath59 is given by @xmath364 where @xmath365 is the off - diagonal entry of the tail - dependence matrix @xmath162 of @xmath281 .
let us now come back to problem ( [ eq : gammad ] ) which motivated our research on tail - dependence matrices . from a practical point of view
, the question is whether it is possible to find one financial position , which has tail - dependence coefficient @xmath212 with each of @xmath366 tail - independent financial risks ( assets ) .
such a construction can be interesting for risk management purposes , for example , in the context of hedging.=-1 recall problem ( [ eq : gammad ] ) : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for which @xmath8 $ ] is the matrix @xmath367 a matrix of pairwise ( either lower or upper ) tail - dependence coefficients ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ based on the frchet
hoeffding bounds , it follows from joe @xcite , theorem 3.14 , that for @xmath11 ( and thus also @xmath12 ) , @xmath212 has to be in @xmath368 $ ] ; however , this is not a sufficient condition for @xmath369 to be a tail - dependence matrix .
the following proposition not only gives an answer to ( [ eq : gammadrec ] ) by providing necessary and sufficient such conditions , but also provides , by its proof , a compatible model for @xmath369 .
[ prop : example ] @xmath370 if and only if @xmath371 .
the if - part directly follows from corollary [ coro : ddm ] .
we provide a constructive proof based on theorem [ thm : main : tdm : y ] .
suppose that @xmath371 .
take a partition @xmath372 of the sample space @xmath373 with @xmath374 , @xmath375 , and let @xmath376 .
it is straightforward to see that=-1 @xmath377=\pmatrix { \alpha & 0 & \cdots & 0 & \alpha\vspace*{2pt } \cr 0 & \alpha & \cdots & 0 & \alpha\vspace*{2pt } \cr \vdots & \vdots & \ddots & \vdots&\vdots \vspace*{2pt } \cr 0 & 0 & \cdots & \alpha & \alpha \vspace*{2pt } \cr \alpha & \alpha & \cdots & \alpha & 1 } .\ ] ] by proposition [ coro - b ] , @xmath378)\in{\mathcal{t}}_d$ ] .
for the only if part , suppose that @xmath370 ; thus @xmath379 . by theorem [ thm : main :
characterization ] , @xmath380 . by the definition of @xmath129 ,
@xmath381 for some @xmath382 $ ] and a bernoulli - compatible matrix @xmath383 .
therefore , @xmath384 is a compatible bernoulli matrix , so @xmath385 .
write @xmath386 $ ] for some @xmath387 .
it follows that @xmath388 for @xmath389 , @xmath390 for @xmath54 , @xmath391 and @xmath392 for @xmath375 .
note that @xmath393 , @xmath375 , are almost surely disjoint since @xmath390 for @xmath54 , @xmath391 . as a consequence , @xmath394 and thus @xmath395 .
it follows from the proof of theorem [ thm : main : tdm : y ] that for @xmath396 $ ] , a compatible copula model with tail - dependence matrix @xmath369 can be constructed as follows .
consider a partition @xmath372 of the sample space @xmath373 with @xmath374 , @xmath375 , and let @xmath397 ; note that @xmath321 here .
furthermore , let @xmath262 be as in theorem [ thm : main : tdm : y ] , @xmath398 $ ] and @xmath399 be independent
. then @xmath400 has tail - dependence matrix @xmath369 .
example [ ex : special : cases ] , part 1 provides the copula @xmath134 of @xmath59 in this case .
it is also straightforward to verify from this copula that @xmath59 has tail - dependence matrix @xmath369 .
figure [ fig : gamma ] displays pairs plots of 2000 realizations of @xmath59 for @xmath401 and two different copulas for @xmath262 . which produces the tail dependence matrix @xmath402 as given by ( [ eq : gammad ] ) . on the left - hand side , @xmath361
[ @xmath212 determines how much weight is on the diagonal for pairs with one component being @xmath403 ; see ( [ ex : special : cases : fd ] ) ] and on the right - hand side , @xmath262 follows a gauss copula with parameter chosen such that kendall s tau equals 0.8 . ] note that @xmath369 is not positive semidefinite if and only if @xmath404 .
for @xmath405 , element - wise nonnegative and positive semidefinite matrices are completely positive ; see berman and shaked - monderer @xcite , theorem 2.4 .
therefore , @xmath406 is completely positive .
however , it is not in @xmath407 .
it indeed shows that the class of completely positive matrices with diagonal entries being 1 is strictly larger than @xmath142 .
inspired by the question whether a given matrix in @xmath10^{d\times d}$ ] is the matrix of pairwise tail - dependence coefficients of a @xmath1-dimensional random vector , we introduced the tail - dependence compatibility problem .
it turns out that this problem is closely related to the bernoulli - compatibility problem which we also addressed in this paper and which asks when a given matrix in @xmath10^{d\times d}$ ] is a bernoulli - compatible matrix ( see question [ q : b : compatibility ] and theorem [ thm : bern ] ) . as a main finding , we characterized tail - dependence matrices as precisely those square matrices with diagonal entries being 1 which are bernoulli - compatible matrices multiplied by a constant ( see question [ q : characterization ] and theorem [ thm : main : characterization ] ) . furthermore , we presented and studied new models ( see , e.g. , question [ q : model : for : y ] and theorem [ thm : main : tdm : y ] ) which provide answers to several questions related to the tail - dependence compatibility problem .
the study of compatibility of tail - dependence matrices is mathematically different from that of covariances matrices . through many technical arguments in this paper
, the reader may have already realized that the tail - dependence matrix lacks a linear structure which is essential to covariance matrices based on tools from linear algebra .
for instance , let @xmath28 be a @xmath1-random vector with covariance matrix @xmath282 and tail - dependence matrix @xmath162 , and @xmath19 be an @xmath408 matrix .
the covariance matrix of @xmath409 is simply given by @xmath410 ; however , the tail - dependence matrix of @xmath409 is generally not explicit ( see remark [ transition ] for special cases ) .
this lack of linearity can also help to understand why tail - dependence matrices are realized by models based on bernoulli vectors as we have seen in this paper , in contrast to covariance matrices which are naturally realized by gaussian ( or generally , elliptical ) random vectors .
the latter have a linear structure , whereas bernoulli vectors do not .
it is not surprising that most classical techniques in linear algebra such as matrix decomposition , diagonalization , ranks , inverses and determinants are not very helpful for studying the compatibility problems we address in this paper .
concerning future research , an interesting open question is how one can ( theoretically or numerically ) determine whether a given arbitrary nonnegative , square matrix is a tail - dependence or bernoulli - compatible matrix . to the best of our knowledge there are no corresponding algorithms available .
another open question concerns the compatibility of other matrices of pairwise measures of association such as rank - correlation measures ( e.g. , spearman s rho or kendall s tau ) ; see @xcite , section 6.2 .
recently , @xcite and @xcite studied the concept of tail - dependence functions of stochastic processes .
similar results to some of our findings were found in the context of max - stable processes . from a practitioner s point - of - view , it is important to point out limitations of using tail - dependence matrices in quantitative risk management and other applications .
one possible such limitation is the statistical estimation of tail - dependence matrices since , as limits , estimating tail dependence coefficients from data is nontrivial ( and typically more complicated than estimation in the body of a bivariate distribution ) . after presenting the results of our paper at the conferences `` recent developments in dependence modelling with applications in finance and insurance2nd edition , brussels ,
may 29 , 2015 '' and `` the 9th international conference on extreme value analysis , ann arbor , june 1519 , 2015 , '' the references @xcite and @xcite were brought to our attention ( see also acknowledgments below ) . in these papers ,
a very related problem is treated , be it from a different , more theoretical angle , mainly based on the theory of max - stable and tawn
molchanov processes as well as results for convex - polytopes . for instance
, our theorem [ thm : main : characterization ] is similar to theorem 6(c ) in @xcite .
we would like to thank federico degen for raising this interesting question concerning tail - dependence matrices and janusz milek for relevant discussions .
we would also like to thank the editor , two anonymous referees , tiantian mao , don mcleish , johan segers , kirstin strokorb and yuwei zhao for valuable input on an early version of the paper .
embrechts , p. , mcneil , a. and straumann , d. ( 2002 ) . correlation and dependence in risk management : properties and pitfalls . in : _
risk management : value at risk and beyond _ , ed . m.a.h .
dempster , cambridge university press , cambridge , 176223 . | the tail - dependence compatibility problem is introduced .
it raises the question whether a given @xmath0-matrix of entries in the unit interval is the matrix of pairwise tail - dependence coefficients of a @xmath1-dimensional random vector .
the problem is studied together with bernoulli - compatible matrices , that is , matrices which are expectations of outer products of random vectors with bernoulli margins .
we show that a square matrix with diagonal entries being 1 is a tail - dependence matrix if and only if it is a bernoulli - compatible matrix multiplied by a constant .
we introduce new copula models to construct tail - dependence matrices , including commonly used matrices in statistics .
./style / arxiv - general.cfg , |
power variability is one of the main obstacles facing renewable energy expansion .
solar and wind are of main concern since their electrical production is proportional to the variability associated with wind and solar resources@xcite .
this poses an issue for current electrical infrastructure which was designed around variability in demand , not generation .
transient local demand changes over seconds or minutes are for the most part small and spatially uncorrelated resulting a relatively steady demand profile . over several hours
, loads can change substantially , but these changes in load have a tendency to be more predictable .
this is manifested through daily patterns of morning load pickup and evening load drop - off highly correlated with human activity@xcite .
wind and solar generation , on the other hand , is variable .
an individual wind turbine or solar plant can ramp from full to less than half of production in a minute .
on the other hand the aggregate variability of multiple turbines at the same site or even all renewable generators in a balancing area is relatively much less @xcite .
still such variability is not entirely predictable and therefore causes uncertainty in projecting power output minutes to days ahead .
variability and uncertainty are more critical in standalone or island mode applications where a high penetration of renewable power sources ramping near synchronously may create power variability that is large enough to cause substantial power quality and/or grid economics issues .
an approach to solve this problem is to increase scheduling and adjustments to controllable loads to `` load follow '' wind and solar generation on the grid @xcite .
examples of controllable loads are devices such as air conditioners and refrigerators with a temperature dead band that effectively creates a thermal storage reservoir .
the load power of these type of machines can be changed temporarily while respecting the demands of the end user @xcite .
another example is scheduling of more intermittent loads such as water pumps , which can be adjusted to accommodate power variability , claim optimal power usage , and decrease power losses . the control and scheduling of such loads benefits from supply and load forecast @xcite .
+ load scheduling has been applied in many fields , such as thermal loads , residential appliances , and ev charging @xcite . in @xcite , a case study
was implemented to accomodate wind power variability through ev charging .
another example for household appliances scheduling was demonstrated in @xcite . from the supply side
, @xcite discussed new dispatch methodology to power and control a hybrid wind turbine and battery system . in @xcite ,
more work was done toward game theory and customers effect on the grid .
+ model predictive control ( mpc ) was used in most approaches@xcite to compute optimal control or scheduling signals for the load .
typically , in mpc a constraint ( quadratic ) optimization problem is solved iteratively over a finite @xmath0 and moving time @xmath1 horizon from @xmath2 till @xmath3 to compute an optimal control signal in real time , denoted here by @xmath4 at time @xmath2 .
countless examples of innovative mpc based approaches for load scheduling , grid tied storage systems or to maintain voltage stability can be found in e.g. @xcite , @xcite , @xcite or @xcite .
although mpc approaches are extremely powerful in computing optimal control signals over a moving but finite time horizon , typically the control signal @xmath4 is allowed to attain any real value during the optimization , see e.g. @xcite .
unfortunately , a real - valued control signal @xmath4 would require distinctive loads on an electric grid to operate at fractional load demands .
although this can be implemented by electric storage systems or partial or pulse width modulation of loads @xcite , ( non - linear ) dynamic power profiles of the electric loads in terms of dynamic power ramp up / down and minimum time on / off of each load is harder to implement in a standard mpc framework .
+ in this paper we define load scheduling as the optimal on / off combinations and timing of a set of distinct electric loads via the computation of an optimal binary control signal @xmath5 . the work is partially motivated by previous work @xcite and @xcite in which the design and sizing problem of a standalone photovoltaic reverse osmosis ( ro ) system is considered , where the ro loads are to be scheduled on / off .
the work in @xcite computes the optimal size and number of units for a selected location but lacks the procedure for optimally scheduling dynamic loads .
here we aim to find optimal load scheduling by on / off switching of possibly non - linear dynamics electric loads .
+ the mpc optimization problem becomes untractable for binary load switching because of the exponentiation growth of the binary combinations in the length of the prediction horizon @xmath0 and the number @xmath6 of loads .
constraints on the allowable load switching help to alleviate the combinatorial problem , making mpc optimization with binary switching computationally feasible .
the approach presented in this paper will be illustrated in a simulation study in which each load has its own dynamics for both turning on and shutting off .
solar forecasting data on a partly cloudy morning and a clear afternoon at uc san diego is used to illustrate how loads are scheduled to turn on / off dynamically to track solar power predictions .
the moving horizon nature of forecasts serve as an ideal input to the mpc algorithm . with the finite prediction horizon in mpc
it is crucial to have reliable forecast of power delivery .
however , in reality cloud advection forecasts become less accurate over longer horizons as the cloud dynamics render the basic assumptions of static clouds invalid .
this inaccuracy leads to error in the mpc decision . either an overprediction can cause loads to activate during a period where there is not enough solar energy to meet the demand or an underprediction can prevent loads from being schedule even though energy would have been available .
for this reason , we will investigate mis - scheduling due to forecast errors .
+ section [ sec : sdlm ] introduces the dynamic loads assumptions in terms of dynamic power ramping and on / off time constraints .
section [ sec : dls ] gives the approach for dynamic load scheduling based on power tracking over a moving prediction horizon of @xmath0 points , with an admissible set of binary switching combinations . in section [ sec : aex ] different solar forecast methods are considered and the effect of forecast errors on the load scheduling is investigated .
advective , persistence and perfect forecasts are used as inputs to the load scheduling algorithm to show how forecasting errors can lead to scheduling errors .
the paper is ended by concluding remarks in section [ sec : ccls ] .
we consider a fixed number of @xmath6 loads where the power demand @xmath7 , @xmath8 as a function of time @xmath1 for each load @xmath9 is modeled by a known switched dynamic system .
for the dynamic scheduling of the loads , loads are assumed to be switched `` on '' or `` off '' by a binary switching signal @xmath10 .
each load @xmath9 is also assumed to have a known minimum duration @xmath11 for the `` off '' time of the load when @xmath12 and a minimum duration @xmath13 for the `` on '' time of the load when @xmath14 .
the duration times @xmath15 and @xmath16 avoid unrealistic on / off chattering of the switch signal @xmath17 during load scheduling and limit the number of transitions in @xmath17 over a finite optimization period @xmath18 . this can be also an equipment safety or operational constrains . with the minimum on / off duration times @xmath19 and the finite time period @xmath20 for load switching , on / off switching of a load at time @xmath21 can now be formalized .
special care should be given to turning on loads at @xmath22 close to the final time @xmath23 which is also depends on the equipments or loads type . for the formalization
, the load switching signal @xmath17 will be a combination of an `` on '' signal @xmath24 and an `` off '' signal @xmath25 that both take into account the constraints of minimum on / off duration and the finite time @xmath20 for load switching . as a result ,
the admissible on / off transition signal @xmath26 of a load at time @xmath21 can now be formalized by the switching signal @xmath27 where @xmath28 denotes the most recent ( last ) time stamp at which the load @xmath9 was switched `` off '' , and the opposite goes to @xmath29 @xcite . for the computational results presented in this paper ,
linear first order continuous - time dynamic models will be used to model the dynamics of the power demand of the loads .
it should be pointed out that the computational analysis is not limited to the use of linear first order models , as long as the dynamic models allow the numerical computation of power demand @xmath7 as a function of the switching signal @xmath17 . to allow different dynamics for the time dependent power demands @xmath7 when
the binary switching signal @xmath10 transitions from 0 to 1 ( `` on '' ) or transitions from 1 to 0 ( `` off '' ) , different time constants are used in the first order models .
this allows power demands @xmath7 to be modeled at different rates when switching loads . referring back to the admissible on / off transition signals @xmath30 and @xmath31 respectively in [ eq : wion ]
, the switched linear first order continuous - time dynamic models for a particular load are assumed to be of the form @xmath32 to model the power demand @xmath7 of a load .
same goes to the off signal model but different time constants @xmath33 and @xmath34 are used to model respectively the on / off dynamic switching of the load .
to achieve the optimal switching times @xmath35 of the binary switching signals @xmath17 for each load discretizing the power demand @xmath7 and the optimal switching signal @xmath17 was performed at a time sampling @xmath36 where @xmath37 is the sampling time and @xmath38 is an integer index . to simplify the integer math
, we assume that both the switching times and the minimum on / off duration times are all multiple of the sampling time @xmath37 . with the imposed time discretization
, the switching signal @xmath39 is held constant between subsequent time samples and @xmath40 and @xmath41 .
a zero order hold ( zoh ) discrete - time equivalent of the continuous - time models given earlier in [ eq : odeon ] was used to achieve the computation of @xmath42 , and it is given by @xmath43 @xmath44 for on switching of the load .
the coefficients @xmath45 and @xmath46 in the first order zoh discrete - time equivalent models are fully determined by the time constants @xmath33 , @xmath34 , static load demand @xmath47 and the chosen sampling time @xmath37 . defining the optimization that allows the computation of optimal discrete - time switching signals @xmath48 for the power demand @xmath42 of @xmath6 loads . defining a power tracking error
@xmath49 it is clear that computing optimal @xmath39 will involve a criterion function and possible constraints on @xmath50 and @xmath39 over a ( finite ) time horizon @xmath51 .
choosing @xmath0 to be large , e.g. @xmath52 where @xmath20 is the complete optimization period , results in two major disadvantages.the first one is that the number of possible combinations of the discretized binary switching signal @xmath39 grows exponentially with the number of loads @xmath6 and the number of time steps @xmath0 .
fortunately , this can be significantly reduced by the requirement of minimum on / off duration times @xmath19 for the loads.as mentioned before , this avoids unrealistic on / off chattering of the switch signal @xmath17 during load scheduling and significantly reduces the number of binary load combinations .
the second disadvantage of choosing @xmath0 to be large requires the discrete - time power profile @xmath53 to be available over many time samples to plan for optimal load scheduling , which leads to increasingly suboptimal schedules due to increasing solar forecast errors . with the imposed time discretization given in [ eq : ni ] , and a finite prediction horizon @xmath0 ,
the admissible on / off transition signal in [ eq : wion ] reduces to @xmath54 where @xmath55 now denotes the most recent discrete - time index at which the load @xmath9 was switched `` off '' .
similarly @xmath56 reduces to @xmath57 where @xmath58 denotes the most recent discrete - time index at which the load @xmath9 was switched `` on '' . both signals @xmath59 in [ eq : wiondiscrete ] and @xmath60 in [ eq : wioffdiscrete ]
form a set @xmath61 of binary values for admissible discrete - time switching signals defined by @xmath62 it is beneficial to note that the number of binary elements in the set @xmath61 is always much smaller than @xmath63 due to required minimum number of on / off samples @xmath64 for the loads .
this results shows that constraints on the allowable load switching helps to alleviate the combinatorial problem , making an optimization with binary switching computationally feasible . as an example , consider the case of @xmath65 loads over a power prediction horizon of @xmath66 samples . without any requirements on minimum number of on / off samples one would have to evaluate @xmath67 possible combinations of the load switching signal @xmath68 .
starting at a binary combination with all loads off , e.g @xmath69 $ ] and requiring the loads to stay on / off for at least 4 samples reduces the number of possible binary combinations to only @xmath70 .
clearly , the number of combinations reduces even further for a non - zero initial condition , e.g. @xmath71 $ ] , where the first load that is switched on is required to stay on over the prediction horizon .
it is clear from the above illustrations that the number of admissible binary combinations of @xmath6 loads over a prediction horizon of @xmath0 points is in general much smaller than @xmath63 , making the optimization with binary switching computationally feasible for real - time operation .
the dynamic load scheduling optimization problem is formulated as a moving horizon optimization problem by following the power tracking error defined in [ eq : powererror ] @xmath72 where @xmath73 + + with the admissible set @xmath61 defined in [ eq : wset ] . adopting the ideas from mpc ,
the @xmath74 dimensional optimal switching signal @xmath75 over the optimization horizon @xmath76 and the loads @xmath8 is selected by the evaluation of the criterion function @xmath77 as a function of the power tracking error @xmath78 ( in the future ) at @xmath79 . once the optimal switching signal @xmath80 ,
@xmath76 is computed , the optimal signal is applied to the loads _ only _ at the time instant @xmath40 , after which the time index @xmath81 is incremented and the optimization in [ eq : mpc ] is recomputed over the moving time horizon . as the admissible set @xmath61 defined in [ eq : wset ] has a finite and countable number of binary combinations for the switching signal ,
the optimal value for @xmath80 is computed simply by a finite number of evaluation of the criterion function @xmath77 .
hence , no ( gradient ) based optimization is used to compute the final value for @xmath39 .
possible candidate functions @xmath77 may include a least squares criterion or may include a barrier function @xmath82 to enforce a positive constraint @xmath83 . such constraints may be required to guarantee that the load demand is always smaller then the ( predicted ) power profile @xmath53 in [ eq : powererror ] . in this paper
we use the quadratic function with a barrier function in [ eq : optcrit ] to perform tracking of predicted solar power curves by dynamic load switching .
this algorithm can be implemented for any kind of standalone system ( wind , solar or even hybrid ) with a forecast tool providing input data . here
, we present a standalone solar system powering 3 units of normalized sizes rated at @xmath84 60% , 26% and 12% of full solar power . furthermore , every load has different dynamics for on / off switching modeled by the first order time constants @xmath33 and @xmath34 similar to the model given in ( [ eq : odeon ] ) .
the first order time constants @xmath33 and @xmath34 are dependent on the size @xmath47 of the load .
as illustrated in figure 3 in @xcite , the proposed scheduling approach schedules the on / off status of the three different loads in order to capture as much solar energy as possible , _
i.e. _ to decrease the unutilized ( lost ) energy
. the algorithm is tested against a clear sky model predicted day as well as the real pv forecast recorded on september 09 , 2014 by ucsd sky imager .
the solar production from the clear sky model is presented in figure [ sim ] . as determined in @xcite , when short - term variability is small and deterministic ,
a smooth scheduling for @xmath65 loads can be achieved , as observed ( figure [ sim ] ) .
it can be seen that the scheduling algorithm emphasizes that turning on the largest unit is the main priority .
lost energy is minimized by combining all three loads through out the day .
seconds and @xmath85 for three loads .
http://solar.ucsd.edu/c/wp-content/uploads/2015/11/bellcurvesim1.gif[animation can be found here . ] ] table [ forecastscenarios ] compares the efficiency of the system , defined as the percentage of solar power the loads capture , under the different forecast scenarios .
a diminishing returns effect is observed for different forecast horizon @xmath0 and switching time @xmath40 . as observed for @xmath86 , the efficiency of the system increases as @xmath40 decreases . for @xmath87 , increases with @xmath0 , but remains constant for @xmath88 . from table
ii we conclude that increasing @xmath0 is computationally costly and provides diminishing returns for @xmath0 greater than 8 times the switch time .
.loads characteristics ( seconds ) [ cols="<,<,<,<,<,<",options="header " , ]
solar forecasting is critical for load scheduling in standalone pv systems due variability of solar resources . uc
san diego has developed a ground based sky imager to detect clouds , cloud velocity , and forecast the advection of cloud shadows on the ground over the coming 10 to 20 minutes @xcite .
the imager is composed of an upward facing camera coupled with a fisheye lens to capture a large area of o(10 km2 ) .
the forecasting algorithm uses projected cloud locations coupled with a clear sky index model to predict global horizontal irradiance ( ghi ) over the captured domain .
this is referred to as `` advection forecast '' .
a commonly implemented forecasting method is to assume that the clear sky index _
kt _ at time _
t _ remains constant for the short period of the forecast _
t+n _ , with irradiance increasing at the clear sky value multiplied by _ kt_. this method is called `` persistence '' forecasting
. any more complex forecasting technique can be benchmarked against these persistence forecasts .
persistence and advection forecast methods are associated with erroneous predictions .
persistence errors are due to the deviation from the initial cloud state , which generally increases in time .
advection errors arise mainly from inaccuracy of cloud detection and mis - prediction of cloud formation , evaporation , or deformation .
for this reason , the accuracy of the ucsd solar forecasting algorithm against perfect , and persistence conditions are compared , and their effects on scheduling algorithms is classified .
the aggregate forecast errors for the forecast for this day are : @xmath89 , @xmath90 , @xmath91 , @xmath92 computed by @xcite .
figure [ compare ] illustrates the results of the mpc scheduling algorithm under the three different forecast scenarios . by design the scheduling under the perfect forecast
is error free .
however , under persistence and advection forecast predictions we can see several areas where the load demand is greater than the available solar energy . + an overprediction can cause loads to activate during a period where there is not enough solar energy to meet the demand ( power exceedence or pe ) ; an underprediction can prevent loads from being scheduled even though energy would have been available resulting in lost energy .
two metrics are considered to evaluate the accuracy of the control algorithms .
the first metric is the number of time steps with pe while the second one is called energy exceedence ( ee ) and is equal to integration of pe at the consecutive violated time steps .
the ee provides a reasonable estimate for the size of an energy storage system that would avoid pe .
+ figure [ compare ] depicts that three pe situations happen for the persistence forecast scenario between 9:00 and 11:00 . at 9:10 ,
the intermediate sized unit ( unit 2 ) is in pe for approximately 10 minutes . at 10:20 ,
the smallest unit ( unit3 ) is in a shallow pe in two distinct periods over 10 minutes .
finally at 10:45 , we see the largest unit is in pe for 5 minutes . referring to the advective forecast we the same exceedance scenarios at 10:20 and 10:45 , but the pe at 9:10 is avoided .
+ figure [ overreach ] shows the total ee for the day , the maximum energy of individual consecutive pe events , and the total number of violations for the day as a function of _
we see that the persistence forecast leads to higher exceedance and more violations for all _ k_. increasing _ k _ in general also leads to an increase in exceedance and violations .
if we evaluated solar forecast or control algorithms solely based on pe , then a forecast that is biased small ( or even always zero ) would improve the results as less and/or smaller loads would be scheduled . therefore a different metric such as efficiency is important to observe .
figure [ efficiency ] shows the efficiency under each scenario as a function of _ k_. it is observed that the perfect forecast has the highest efficiency . for low _
k _ , the advective forecast captures more energy , where the opposite is true for high k. + in both figures [ overreach ] and [ efficiency ] , there is non - monotonic behavior associated with increasing _
k _ , which is not expected . for this reason
, the same analysis was run with three units of equal size ( _ @xmath47 _ ) .
this is plotted in figure [ efficiency ] . for this case ,
the efficiency decreases monotonically with _
k _ ( and is much lower than the non - equal case ) , and the advective forecast always captures more energy than the persistence forecast . we can conclude that the nonlinear behavior seen in non - equal units across increasing _
k _ is due to the dynamics associated with starting and shutting down the loads .
+ energy storage systems such as batteries could help overcome ee episodes as well as increase the efficiency .
the excess of energy that is not captured by the loads ( figures [ compare],[efficiency ] ) could be stored and used to power loads during ee events . for battery sizing one could consider match the energy capacity of the battery to the energy required for maximum exceedance of the largest unit over its minimum required on - time .
however , figure [ overreach ] demonstrates that the maximum individual ee for all cases is a small percentage the power of even the smallest unit .
reducing the size of the battery reduces the up front capital cost associated with the system . a more sophisticated method for battery
sizing would be to probabilistically determine the maximum individual ee case for several years of historical forecast data and size accordingly , which will be the focus of future work .
in this paper , a mpc model was developed to compute the optimal binary control signal by determining the on / off combinations and timing of a set of distinct electric dynamic loads scheduling .
the mpc load scheduling algorithm was tested using different forecasting techniques to assess the effects of input inaccuracy .
the algorithm worked optimally under a perfect forecast ( as designed ) , but created errors due to forecasting error .
the energy captured by the loads decreased for increasing _
k _ in a non - monotonic manner for optimally sized unequal units . however , for the same scenario with equally ( but not optimal ) sized units , efficiency decreased monotonically with increasing _
k _ for all forecast scenarios .
the advection forecast model created fewer errors and gave less total exceedance and number of violations independent of the _ k _ , as compared to persistence . for a stand alone system these errors
could be mitigated with storage capacity to provide the power to the load to ride through the exceedance period .
a battery sizing method is discussed as the topic of future work .
wan , y. h. `` long - term wind power variability '' .
nrel / tp-5500 - 53637 .
n.p . , jan .
2012 . web .
makarov , yuri v. , clyde loutan , jian ma , and phillip de mello .
`` operational impacts of wind generation on california power systems . '' power systems , ieee transactions on 24 , no
. 2 ( 2009 ) : 1039 - 1050 .
halamay , douglas , ted ka brreserve requirement impacts of large - scale integration of wind , solar , and ocean wave power generationekken , asher simmons , and shaun mcarthur . `` .
'' sustainable energy , ieee transactions on 2 , no . 3 ( 2011 )
: 321 - 328 .
mathieu , johanna l. , maryam kamgarpour , john lygeros , goran andersson , and duncan s. callaway . ``
arbitraging intraday wholesale energy market prices with aggregations of thermostatic loads . ''
power systems , ieee transactions on 30 , no . 2 ( 2015 ) : 763 - 772 .
makarov , yuri v. , pavel v. etingov , jian ma , zhenyu huang , and kamesh subbarao .
`` incorporating uncertainty of wind power generation forecast into power system operation , dispatch , and unit commitment procedures . ''
sustainable energy , ieee transactions on 2 , no
. 4 ( 2011 ) : 433 - 442 .
lin , yashen , prabir barooah , and j. mathieu .
`` ancillary services to the grid from commercial buildings through demand scheduling and control . ''
t. ferhatbe , g. govi zucker , p. palensky , model based predictive control for a solar - thermal system . in africon , 2011 ( pp
. 1 - 6 ) .
ieee , 2011 .
agnetis , alessandro , gianluca de pascale , paolo detti , and antonio vicino .
`` load scheduling for household energy consumption optimization . '' smart grid , ieee transactions on 4 , no . 4 ( 2013 ) : 2364 - 2373 .
kim , byung - gook , shaolei ren , mihaela van der schaar , and jang - won lee . `` bidirectional energy trading and residential load scheduling with electric vehicles in the smart grid . '' selected areas in communications , ieee journal on 31 , no . 7
( 2013 ) : 1219 - 1234 .
morari , manfred , and jay h. lee .
`` model predictive control : past , present and future . ''
computers & chemical engineering 23 , no . 4 ( 1999 )
mayne , david q. , james b. rawlings , christopher v. rao , and pierre om scokaert .
`` constrained model predictive control : stability and optimality . ''
automatica 36 , no . 6 ( 2000 ) : 789 - 814 .
wang , tao , haresh kamath , and steve willard .
`` control and optimization of grid - tied photovoltaic storage systems using model predictive control . ''
smart grid , ieee transactions on 5 , no
. 2 ( 2014 ) : 1010 - 1017 .
abdulelah h habib , vahraz zamani , and jan kleissl .
`` solar desalination system model for sizing of photovoltaic reverse osmosis ( pvro ) . '' in asme 2015 power conference , pp .
american society of mechanical engineers , 2015 .
abdulelah h habib , vahraz zamani , raymond a. de callafon , jan kleissl , `` sizing of photovoltaic reverse osmosis for solar desalination based on historical data in coastal california '' , international desalination association world congress , 2015 .
yang , handa , ben kurtz , dung nguyen , bryan urquhart , chi wai chow , mohamed ghonima , and jan kleissl .
`` solar irradiance forecasting using a ground - based sky imager developed at uc san diego . '' solar energy 103 ( 2014 ) : 502 - 524 . | this paper presents and evaluates the performance of an optimal scheduling algorithm that selects the on / off combinations and timing of a finite set of dynamic electric loads on the basis of short term predictions of the power delivery from a photovoltaic source . in the algorithm for optimal scheduling ,
each load is modeled with a dynamic power profile that may be different for on and off switching .
optimal scheduling is achieved by the evaluation of a user - specified criterion function with possible power constraints .
the scheduling algorithm exploits the use of a moving finite time horizon and the resulting finite number of scheduling combinations to achieve real - time computation of the optimal timing and switching of loads .
the moving time horizon in the proposed optimal scheduling algorithm provides an opportunity to use short term ( time moving ) predictions of solar power based on advection of clouds detected in sky images .
advection , persistence , and perfect forecast scenarios are used as input to the load scheduling algorithm to elucidate the effect of forecast errors on mis - scheduling .
the advection forecast creates less events where the load demand is greater than the available solar energy , as compared to persistence . increasing the decision horizon leads to increasing error and decreased efficiency of the system , measured as the amount of power consumed by the aggregate loads normalized by total solar power . for a standalone system with a real forecast ,
energy reserves are necessary to provide the excess energy required by mis - scheduled loads . a method for battery sizing
is proposed for future work . |
a displacement process involving two fluids is often unstable when the displacing fluid has larger viscosity than the displaced one .
the resulting instability developing at the interface between two fluids is known as viscous fingering @xcite .
this instability has received much attention as an archetype of pattern - formation problems and as a limiting factor in the recovery of crude oil .
classical mathematical model describing a newtonian flow displacement in a hele - shaw cell and development of the saffman
taylor instability consists of the continuity equation , darcy s law and a convection - diffusion equation for the concentration of the displacing fluid @xcite .
the inertia of fluid may be important for high finger velocities .
this leads to the necessity to use more complex nonlinear equations of fluid motion @xcite . in the framework of these models instability caused by different velocities of layers movement
can be considered .
there are several theoretical and experimental studies on the role of inertia in immiscible @xcite and miscible @xcite displacements .
the results reveal that inertia tends to damping viscous fingering . in recent publications @xcite the effects of the thickness variation of a hele - shaw cell and elasticity of the walls on the process of viscous fingering
have been studied .
different types of instability in viscosity - stratified flow have been discussed in @xcite .
a number of theoretical , numerical , and experimental works devoted to understanding of various aspects of the instabilities is available in the literature @xcite .
accurate numerical solution is costly at high peclet numbers and it is difficult to reproduce the detailed fingering pattern .
numerical simulations @xcite show that it is simpler to describe the concentration of solute averaged across the fingers . in such cases ,
the mixing zone is an important feature to determine the extent of mixing . despite the considerable work done
, the spreading and growth of the mixing zone is an important question that still remains unresolved .
several empirical models are available for the evaluation of mixing zones in unstable , miscible displacements .
two empirical models have been suggested by koval @xcite and todd and longstaff @xcite to give a basis for computation of miscible displacement .
both models suffer from adoption of empiricism in which the principal parameters involved have little or indirect physical significance .
further development of averaged models of fingers formation is represented in @xcite .
all these models are based on the hypothesis of pressure equalization in the transverse direction to the main flow , as well as an empirical information about the displacing and displaced fluids distribution in the region of intensive viscous fingering .
let us recall that 2d numerical solution with high resolution is hard to construct , that is why 1d models play an important role in some cases .
for instance , these models are very useful in calculating of fracturing when it is necessary to solve the equations of the crack opening and the fluid motion in the fracture simultaneously @xcite .
the aim of the present paper is to derive a hierarchy of mathematical models describing viscosity - stratified flow and spreading and growth of the mixing zone in a hele - shaw cell . in section 2
we propose 2d nonlinear hyperbolic system of balance laws .
in contrast with widespread model of flow displacement in a hele - shaw cell we apply nonlinear momentum equations and take into account compressibility of the fluid . at the same time diffusion coefficient
is neglected that corresponds to the large peclet number limit . as we show in section 3 by numerical calculations ,
this model is suitable for describing of jet flow and propagation of viscous fingers in a hele - shaw cell .
we also point out that for the process of unidirectional displacement the pressure variation in the transverse direction is small .
this observation makes it possible to use long - wave approximation and construct a class of layered flows described by a system of one - dimensional evolution equations .
based on the various simplifications of the momentum equation ( linearisation , lubrication theory ) a hierarchy of 1d mathematical models is constructed in section 4 .
the equations of a three - layer flow are studied and numerical computations of the formation of viscous fingers are performed in section 5 .
we show that three - layer stationary flow is correctly described in the framework of simplified model .
nevertheless the growth rate of viscous fingers is significantly higher than it is observed experimentally . in section 6
we propose 1d kinematic - wave model of viscosity - stratified flow taking into account friction between the fluid layers .
the velocity of propagation and the thickness of the viscous finger in the framework of this kinematic - wave model coincide with the corresponding calculations on the basis of the 2d equations .
it gives possibility to predict the parameters of viscous fingers without time - consuming calculations .
we also show that the proposed 1d model is in good agreement with the well - known koval model .
a newtonian weakly - compressible flow displacement in a hele - shaw cell ( the area between two parallel plates separated by a small gap of constant thickness @xmath3 in the @xmath4 direction , see fig .
[ fig : fig_1 ] ) is described by the equations @xmath5 here @xmath6 is the fluid velocity , @xmath7 is the density , and @xmath8 is the stress tensor .
this tensor can be taken in the form @xmath9 , where @xmath10 is the pressure , and @xmath11 is the viscosity . to describe the process of displacement involving two fluids of different constant viscosities we should take into account that viscosity @xmath11 depends on the concentration of solvent @xmath12 .
this function is scaled such that it is equal to unity in the displaced fluid ( @xmath13 ) and zero in the displacing one ( @xmath14 ) . following @xcite we assume a monotonic relationship between the viscosity and the concentration in the form @xmath15 .
the concentration @xmath12 satisfies to the transport equation ( diffusion is neglected that corresponds to the large peclet number limit ) @xmath16 we assume now that the velocity field can be represented as @xmath17 it provides the fulfilment of no - slip conditions on the cell walls @xmath18 .
we also suppose that the functions @xmath10 , @xmath7 , and @xmath12 do not depend on @xmath4 .
let us note that in the calculation of @xmath19 the following terms @xmath20 , @xmath21 , and @xmath22 can be omitted as they are negligible compared to the derivatives with respect to @xmath4 .
further , averaging eqs .
and through the gap we obtain ( primes are omitted ) @xmath23\displaystyle ( \rho v)_t+(\beta\rho uv)_x+(\beta\rho v^2+p)_y=-\mu v , \\[2mm]\displaystyle \rho_t+(u\rho)_x+(v\rho)_y=0 , \quad ( c\rho)_t+(uc\rho)_x+(vc\rho)_y=0 .
\end{array}\ ] ] here and below @xmath11 denotes the viscosity of the fluid divided by the permeability @xmath24 ( further in the text we will call it simply `` viscosity '' ) ; coefficient @xmath25 is equal @xmath26 ( this factor comes from integration of @xmath27 in the form with respect to @xmath4 @xcite ) . in order to close model we should specify either the equation of state @xmath28 ( barotropic fluid ) or dependence @xmath29 ( incompressible fluid ) .
a weak compressibility of the fluid given by the equation of state @xmath28 provides hyperbolicity of the model .
this dependence can be considered as a regularization of equations describing the flow of an incompressible fluid in a hele - shaw cell . on the other hand in applications
the property of hyperbolicity of equations can be related to the presence of gas cavities in the porous medium as well as with the elasticity of the channel walls ( for instance in pkn model @xcite the pressure @xmath10 depends on the channel thickness @xmath3 ) . in any case
if the condition @xmath30 holds then results weakly depend on the choice of @xmath28 .
therefore for the numerical simulation of 2d flows we assume @xmath31 where the constants @xmath32 and @xmath33 specify characteristic density and speed of sound in the fluid . to find the characteristics of system , we write it in the vector form @xmath34 where @xmath35 is the vector of dependent variables ; @xmath36 is the right - hand side ; @xmath37 and @xmath38 are @xmath39 matrices .
let @xmath40 be the normal vector to the characteristics ; @xmath41 is the identity matrix .
then the characteristic matrix @xmath42 of system , has the form @xmath43 ( \beta-1)v\xi_2 & \chi_1+(\beta-1)v\xi_3 & \big((\beta-1)(\chi_2-\xi_1)v+p'(\rho)\xi_3\big)\rho^{-1 } & 0 \\[1 mm ] \rho\xi_2 & \rho\xi_3 & \chi_2 & 0 \\[1 mm ] 0 & 0 & 0 & \chi_2 \end{pmatrix } \,.\ ] ] here @xmath44 , @xmath45 .
a simple but cumbersome calculation yields the following expression for @xmath46 : @xmath47 we specify the characteristic surface by the equation @xmath48 . then to obtain the differential equations of the characteristics we should replace the vector @xmath49 in previous equation by the vector @xmath50 and equate to zero @xmath51 . as a result
we obtain two families of contact characteristics @xmath52 and two additional characteristic families @xmath53 if the inequalities @xmath54 and @xmath55 hold this system of equations is hyperbolic .
note that in the case of @xmath56 , @xmath57 , and @xmath58 system , coincides with the well - known shallow water equations .
below we present the results of numerical calculations of the viscous fingers and jet streams on the basis of hyperbolic model ,
. originally mathematical description of the saffman
taylor instability was given in the framework of the darcy s law and the mass conservation equation @xcite .
the inertia of the fluid may be significant for high finger velocities . in @xcite simulation of viscous fingers
was performed using non - linear equations of an incompressible fluid .
we show that eqs . , taking into account the forces of inertia and compressibility of the fluid could be also used for description of this instability . to solve differential balance laws ,
numerically one can apply methods based on various modifications of godunov s scheme . in this work
we implement the robust and stable nessyahu
tadmor second - order central scheme @xcite . in every test
we assume that on the boundaries @xmath59 and @xmath60 the impermeability condition @xmath61 is fulfilled .
the size of the computational domain is @xmath62 , @xmath63 ; the resolution of the problem on the @xmath64 and @xmath65 axes are 300 and 150 nodes correspondingly ( uniform grid ) .
we assume that @xmath66 , @xmath67 and @xmath68 ( for the third test @xmath56 ) .
the values of the variables are considered as dimensionless .
below we present calculations showing the possibility of modelling the evolution of perturbations caused by the kelvin helmholtz and/or saffman taylor instabilities on the basis of the hyperbolic model , .
let at the initial time @xmath69 the hele - shaw cell be occupied by a quiescent fluid having density @xmath70 and viscosity @xmath71 . through the left central cross - section of width @xmath72 fluid of viscosity @xmath73
is injected with velocity @xmath74 ; through the rest part of the left boundary fluid of viscosity @xmath71 is entered with velocity @xmath75 . on the right boundary of the domain
the condition of constant pressure is valid . for more intensive development of the perturbations at each time step we slightly disturb boundary conditions at @xmath76 .
namely , the cross section , through which fluid `` 2 '' is injected , is randomly shifted up / down from its initial position on fixed distance @xmath77 which is equal to the grid spacing with some positive integer factor @xmath78 .
we call it `` random shake '' and take here @xmath79 . the function @xmath12 for values of the concentration
is presented in fig .
[ fig : fig_2 ] at @xmath80 .
vortices are formed at the interface of the layers due to kelvin
helmholtz instability .
increase of the both values @xmath81 and @xmath82 suppresses this instability .
+ viscosity - stratified multi - layer flow is shown in fig .
[ fig : fig_3 ] at @xmath83 .
initially the flow region is filled by a quiescent fluid ( @xmath84 , @xmath70 ) .
liquid is pumped through the left boundary divided into layers of height @xmath85 ; velocity and viscosity for odd and even layers are @xmath86 , @xmath87 and @xmath88 , @xmath84 , correspondingly . as in the previous example
`` random shake '' of the jets is used .
the results of the calculations show that multi - layer flow without mixing is realized for a wide range of parameters ( kelvin helmholtz instability at the interfaces between the layers occurs if viscosity decreases more than in five times ) .
we note that the condition on the left boundary @xmath89 corresponds to a class of exact solutions of equations for an incompressible fluid : @xmath90 , @xmath61 , @xmath91 , @xmath92 , @xmath93 .
the stability analysis of this class of flows was carried out in @xcite .
the following example illustrates the formation of viscous fingers .
let the displacing phase injected at a constant velocity @xmath94 be referred to with index 1 and the displaced one with index 2 . at @xmath69 fluid
`` 1 '' is located in the domain @xmath95 ; more viscous fluid `` 2 '' in the domain @xmath96 . for convenience
we use the coordinate system moving with velocity @xmath94 and assume @xmath56 ( with @xmath68 results are similar ) .
let us perturb the initial interface @xmath97 as follows : @xmath98 ( here @xmath99 ) .
we note that the physical effect of the instability can be obtained numerically if the initial perturbation is not less than the grid resolution . at @xmath69
we choose piece - wise linear pressure distribution ( @xmath100 for @xmath101 and @xmath102 for @xmath96 , on the right boundary @xmath10 is equal to @xmath103 ) .
initial density of the fluid is determined using formula . at the boundaries
the impermeability condition is fulfilled .
the calculations are performed for the following parameters : @xmath104 , @xmath84 , and @xmath105 . in the evolution process of the flow
viscous fingers are formed ( fig .
[ fig : fig_4 ] , left ) .
the number of fingers is determined by the initial perturbation .
displacing fluid penetrates more rapidly into displaced one ( fingers are not symmetrical with respect to the initial interface ) .
[ fig : fig_4 ] ( right ) shows the distribution of the density at @xmath106 .
calculation with better resolution leads to the same result . as we can see the density changes less than @xmath107 in comparison with the initial one ( to reduce the compressibility we should increase the speed of sound @xmath33 but this slows down the calculations since the time step is determined by the courant number ) .
note that the density ( pressure ) varies slightly with respect to @xmath65 .
this allows one to use approximate model , where the second momentum equation is replaced by @xmath108 .
we consider an incompressible ( @xmath70 ) hele - shaw flow in a two - dimensional domain of rectangular geometry ( dimensions @xmath109 and @xmath110 , respectively ) governed by eqs . .
it is assumed that the flow is essentially parallel and the pressure gradient in the flow direction @xmath64 being independent of the transverse coordinate @xmath65 to leading - order in the variable @xmath111 .
this regime is termed as the state of transverse flow equilibrium @xcite .
such flows can be also considered in the framework of long wave approximation @xcite .
let us perform the following scaling in eqs .
@xmath112 then we neglect terms of order @xmath113 . as a result
we obtain the approximate model @xmath114\displaystyle u_x+v_y=0 , \quad c_t+uc_x+vc_y=0 , \\[2mm]\displaystyle v\big|_{y=0}=0 , \quad v\big|_{y = h}=0 , \end{array}\ ] ] where the pressure does not depend on the variable @xmath65 .
we also suppose that on the boundaries @xmath59 and @xmath60 the impermeability condition is fulfilled .
let us consider the class of viscosity - stratified flows @xmath115 @xmath116 .
in this case eqs .
take the form @xmath117\displaystyle \sum\limits_{i=1}^n h_i = h , \quad \sum\limits_{i=1}^n u_i h_i = q .
\end{array}\ ] ] here @xmath118 is the depth of @xmath119th liquid layer of viscosity @xmath120 having velocity @xmath121 ; and @xmath122 is the total flow rate through the cell . upon derivation of eqs .
the kinematic condition at the layers interface is used .
introducing new unknown variables @xmath123 allows to transform eqs . to the evolution system of @xmath124 equations @xmath125\displaystyle h_{it}+((s_i+u_n)h_i)_x=0 , \quad\quad ( i=1, ...
,n-1 ) \end{array}\ ] ] where @xmath126 further we assume that @xmath127 .
we also use the following simplified versions of governing eqs . .
the first one consists in the linearisation of the momentum equations : @xmath128 ( here @xmath129 is the average velocity ) . the second simplification is based on the darcy law : @xmath130 the remaining equations of system do not vary . in some cases
it is convenient to use a moving coordinate system @xmath131 , @xmath132 .
then eqs .
take the form ( primes are omitted ) @xmath133\displaystyle h_{it}+(u_i h_i)_x=0 , \quad \sum\limits_{i=1}^n h_i = h , \quad \sum\limits_{i=1}^n u_i h_i=0 . \end{array}\ ] ] further we show that in the framework of three - layer and two - layer regimes of flow it is possible to give an interpretation of the saffman
taylor instability as well as to describe the initial stage of viscous fingering .
we introduce the following notation for the layer velocities and depths @xmath134 we also assume that @xmath135 , @xmath136 , @xmath137 , and @xmath138 . here
we construct a steady - state solution of model for three - layer flow .
integration of the equations of conservation of mass in system allows to express the depths of the layers @xmath139 here @xmath140 is the flow rate in the @xmath141-th layer ( @xmath142 ) . due to the unit depth
we obtain the velocity in the intermediate layer @xmath143 eliminating pressure @xmath10 from the equations @xmath144 ( here the prime denotes the derivative with respect to @xmath64 ) reduces the problem to the solution of the autonomous system of ordinary differential equations @xmath145 a fixed point of system is determined from the relations @xmath146 : @xmath147 linearisation of eqs . on the solution @xmath148 , @xmath149 and computation of the eigenvalues of the corresponding matrix show that the fixed point is a stable node .
the integral curves in the phase plane @xmath150 in the neighbourhood of the fixed point are shown in fig .
[ fig : fig_5 ] for @xmath151 , @xmath152 , @xmath153 , and @xmath68 .
+ fig .
[ fig : fig_6 ] shows a comparison of the numerical results obtained on the basis of two - dimensional hyperbolic equations , and multilayer model reduced to dynamical system in the case of three - layer stationary flow .
solid white lines indicate the layers depths @xmath154 and @xmath155 obtained by solving equation and using relations ; dotted lines correspond to the fixed point ( @xmath156 , @xmath157 ) . here
we take the following values of layers depths @xmath158 , @xmath159 at @xmath76 .
as before we choose @xmath151 , @xmath152 , @xmath153 , and @xmath68 .
the figure shows that the solution reaches an equilibrium state for @xmath160 . to carry out the calculation on the basis of 2d equations ,
the following initial data are used . the flow domain in the @xmath65-direction
is divided into three layers of width @xmath161 , @xmath162 , and @xmath163 .
the fluid of density @xmath70 at @xmath69 moves in these layers in the @xmath64-direction with constant velocities @xmath164 , @xmath165 , and @xmath166 respectively
. we also suppose that @xmath167 in the layers of width @xmath161 and @xmath163 ; in the middle layer of width @xmath162 we choose @xmath153 .
the same data are taken as the boundary conditions at @xmath76 ; on the right boundary ( @xmath168 ) the condition of constant pressure is prescribed ; on the walls @xmath59 and @xmath169 the condition of impermeability is fulfilled .
the resolution of the problem on the @xmath64 and @xmath65 axes are 300 and 60 nodes correspondingly ( uniform grid ) . in order to visualize the flow of fluids with different viscosities the concentration @xmath12 is used .
this value is presented in fig .
[ fig : fig_6 ] at @xmath170 .
more viscous fluid in the middle layer is shown in brown ( @xmath171 ) and less viscous one is shown in blue ( @xmath172 ) .
let us consider a three - layer flow governing by eqs .
wherein the momentum equations are replaced by linear darcy laws @xmath173 .
taking into account assumptions above and notations we have @xmath174 in this case the depths of the layers @xmath175 and @xmath176 are found from the system of equations @xmath177 it is easy to check that this system is hyperbolic and its characteristic velocities are @xmath178 the first family of characteristics is genuinely nonlinear whereas the second one is linearly degenerate @xcite . in terms of the riemann invariants @xmath176 and @xmath179 eqs .
take the form @xmath180 in the case @xmath181 we construct a centred simple wave solution defined by the relations @xmath182 ( note that the ansatz @xmath183 leads to a constant solution ) .
the layer depths are @xmath184 formulae give the solution of eqs . with discontinuous initial data
@xmath185 ( h_0 , \ 0 ) , & \quad x > x_0 . \end{array } \right.\ ] ] profiles of a viscous finger @xmath186 and @xmath187 given by are shown in fig .
[ fig : fig_7 ] for @xmath188 and various values of @xmath189 .
let us construct a solution of cauchy problem , for @xmath190 . at initial time the velocities and the layers
depths are @xmath191 , @xmath192 , @xmath193 , @xmath194 ; @xmath195 , @xmath196 , @xmath197 ,
it is easy to verify that these values satisfy the hugoniot conditions @xmath199=0 , \quad [ ( v - d)\eta]=0\ ] ] derived from eqs . as well as the stability conditions @xcite if the shock front moves with average flow velocity @xmath200 .
it is interested to note that in the class of simple wave solutions ( @xmath201 ) system reduces to the nave koval model @xcite .
in fact , taking into account representation the second equation in can be rewritten as @xmath202 where @xmath203 stands for @xmath176 and @xmath204 .
it is known that the growth rate of viscous fingers in the framework of the nave koval model is significantly higher than it is observed experimentally . to derive another model of a three - layer flow in a moving coordinate system we use linearisation of momentum equations in @xmath205 eliminating the pressure
@xmath10 leads to the system of evolution equations @xmath206\displaystyle s_{2t}+\gamma s_{2x}=-\mu s_2+(1-\mu)(1-w ) , \\[2mm]\displaystyle h_t+(u h)_x=0 , \quad \eta_t+(v\eta)_x=0 \end{array}\ ] ] where @xmath207 + let us rewrite eqs . in the form @xmath208 . here
@xmath209 is the unknown vector , @xmath37 is the jacobi matrix , and @xmath210 is the right - hand part .
the eigenvalues of matrix @xmath37 are @xmath211\displaystyle \lambda_{3,4}=2^{-1}\big((1 - 3h)s_1+(1 - 3\eta)s_2\pm\sqrt{m}\big ) \end{array}\ ] ] where @xmath212 .
conditions @xmath213 and @xmath214 provide hyperbolicity of system since @xmath215 . introducing the parabolic with respect to @xmath216 function @xmath217 it is easy to check validation of inequalities @xmath218 and @xmath219 ( here @xmath220 is a minimum point of @xmath221 ) .
it means that @xmath215 and , consequently , characteristic velocities @xmath222 are real .
further we construct the numerical solution of eqs . with initial data @xmath223 ( 0 , ( 1-\mu)/\mu , h_0 , \ 0 ) , & \quad x > x_0 .
\end{array } \right.\ ] ] note that this formulation corresponds to cauchy problem , .
calculations on the basis of model are carried out using nessyahu
tadmor scheme @xcite .
the results of computations are shown in fig .
[ fig : fig_8 ] at various moments of time ( solid curves ) .
as we can see with increasing of time the solution tends to the self - similar regime ( dot - dashed curves obtained by using formulae ) .
moreover tip of the viscous finger propagates with the same velocity in the framework of models and .
in the previous section it is shown that three - layer flow governed by the simplified 1d model ( or ) correctly describes the well - known fact that the fluid interface is unstable if less viscous fluid displaces more viscous one and stable otherwise . however , the velocity of the viscous fingers propagation for these equations is the same as for the nave koval model which vastly over - predicts the rate of the fingers grow @xcite .
a number of empirical models is proposed to reconcile this behaviour @xcite .
for example , koval @xcite postulates empirically that is valid if the viscosity ratio @xmath224 is replaced by an effective viscosity ratio @xmath225 , where @xmath226 despite the fact that this model appears to work in practice @xcite a theoretical justification is yet to be obtained . as it can be seen from fig .
[ fig : fig_7 ] and [ fig : fig_8 ] , in the framework of model or , the tip of the viscous finger becomes infinitely thin with time .
however , in experiments the formation of fingers of finite thickness is observed .
the friction between the fluid layers is one of the possible mechanisms that prevent thinning of the viscous finger tip .
below we obtain modification of the above considered 1d models , which takes into account friction between the fluid layers .
this model is new and correctly describes growth of the viscous fingers .
it is proved by comparing with the koval model ( eq . where an effective viscosity ratio @xmath225 is used ) as well as by numerical calculations on the basis of 2d eqs . , .
let us consider a two - layer flow governed by the following system of equations @xmath227\displaystyle
v_t+\beta vv_x+p_x = -\mu_2 v+ \kappa\bar{\mu } ( u - v)\eta^{-1 } , \\[2mm]\displaystyle h_t+(uh)_x=0 , \quad \eta_t+(\eta v)_x=0 , \\[2mm]\displaystyle h+\eta=1 , \quad uh+v\eta=1 .
\end{array}\ ] ] here we use notations of the previous section for the layers velocities and depths ; constant @xmath228 is an empirical parameter , and @xmath229 .
this system differs from eqs .
( in the case @xmath230 ) by the presence of additional terms with factor @xmath231 which expresses friction on the layers interface .
note that this effect becomes significant if the thickness of one of the layers tends to zero .
further we use simplification of the momentum equations in based on the darcy law : @xmath232 taking into account that @xmath233 we obtain the following kinematic - wave model @xmath234 where @xmath235 . typical graph of the flux @xmath236 is given in fig .
[ fig : fig_9 ] ( curve _ 1 _ ) for @xmath237 , and @xmath238 . obviously @xmath236 is a non - convex monotonic function such that @xmath239 , and @xmath240 .
we construct self - similar solution of eq . with initial data @xmath241 for @xmath101 and @xmath242 for @xmath96 . on the interval @xmath243 function @xmath236
has two points of inflection .
thus it is necessary to construct convex hull of the flux @xmath236 . for this
we draw tangents to @xmath236 from the origin and from the point @xmath244 ( lines _ 2 _ ans _ 3 _ in fig .
[ fig : fig_9 ] ) .
let @xmath245 and @xmath246 be the tangency points ( values @xmath247 and @xmath248 are obtained by solving the equations @xmath249 and @xmath250 ) .
according to @xcite solution of the cauchy problem takes the form @xmath251
\phi'(h ) , & \quad h \in [ h_1 , h_2 ] \\[2 mm ] \phi'(h_2 ) , & \quad h \in ( h_2 , 1 ] \end{array } \right.\ ] ] where @xmath252 is the self - similar variable .
this means that the solution is presented by centred rarefaction wave , which is bounded by two `` sonic '' shocks . as we show below the numerical results of the growth of viscous fingers in the framework of eqs .
, are in good agreement with the self - similar solution of eq . for a suitable choice of the parameter @xmath253 .
+ let us verify kinematic - wave model by comparing with the well - approved koval model , where instead of @xmath224 effective viscosity ratio @xmath225 given by formula is used .
we choose the following parameters : @xmath254 , @xmath255 .
it means that the viscosity ratio @xmath256 and according to @xmath257 .
self - similar solution of the koval model for these parameters is shown in fig .
[ fig : fig_10 ] ( solid curve ) . corresponding solution of the scalar conservation law ( with @xmath238 ) is presented in fig .
[ fig : fig_10 ] by dashed curve . as we can see in the framework of these models the growth rates of viscous fingers are similar .
although the parameter @xmath253 is a function of @xmath224 , the values of @xmath258 are weakly vary for @xmath259 .
therefore , for the moderate viscosity ratios @xmath224 ( less than 10 ) we can assume that @xmath238 .
moreover , the proposed model describes better the structure of the finger ( its thickness near the tip ) than the koval model .
+ further we compare the results of numerical simulations of the formation of viscous fingers on the basis of 2d hyperbolic system of equations , with the results obtained by using kinematic - wave model .
the calculations are performed in the hele - shaw cell with sizes @xmath260 , @xmath261 in the coordinate system moving with the average flow velocity @xmath262 with respect to @xmath263 axis .
the viscosities of the fluids are equal to @xmath254 and @xmath255 . at the initial time
the interface @xmath264 is perturbed as follows @xmath265 , where @xmath266 .
for discretization with respect to @xmath64 and @xmath65 we use 400 and 50 nodes respectively ( calculation with finer resolution leads to the same result ) . on the left and right boundaries of the computational domain
the reflection conditions are imposed . in 2d model , the following dependence for viscosity is used : @xmath15 .
the calculations are carried out for @xmath56 , @xmath66 and the sound velocity @xmath267 . in this case
the change in density is not more than 0.15% . at the same time
the condition @xmath108 is fulfilled with high accuracy .
it means that the above proposed 1d model is suitable for describing of such flow .
given above perturbation of the interface leads to the formation of a single finger which is symmetric with respect to the line @xmath169 .
the results of the concentration @xmath12 calculations using model , at time @xmath80 and @xmath268 are shown in fig .
[ fig : fig_11 ] .
the distribution of @xmath12 is presented in two colours ( blue for @xmath269 and brown for @xmath270 ) . self - similar solution of kinematic - wave model with @xmath238 is shown in fig .
[ fig : fig_11 ] at @xmath80 and @xmath268 by white solid lines respectively . the curves @xmath271 and @xmath272 for the model are given for a correct comparison with 2d calculations . here
self - similar variable @xmath273 is replaced by @xmath274 .
it corresponds to a transition in a moving coordinate system .
the figure shows the propagation velocity and thickness of the viscous finger obtained by the kinematic - wave model agree well with the two - dimensional calculations .
we derive nonlinear hyperbolic system of equations , describing the flow of slightly compressible multicomponent fluid of different viscosity in a hele - shaw cell . on the basis of these equations simulation of jet flow and development of viscous fingers during the displacement process
are performed .
calculations show that the proposed model reproduces the characteristic features of the flow associated with the development of kelvin helmholtz and saffman taylor instabilities ( see fig .
[ fig : fig_2 ] , [ fig : fig_3 ] and [ fig : fig_4 ] ) . in the case of preferential flow in the @xmath64-direction
the pressure varies slightly in the @xmath65-direction that allows to apply model of long - wave approximation and to consider the class of layered flows described by eqs . .
using various simplifications of the system ( linearisation of the momentum equations , application of the darcy s law ) we construct a hierarchy of mathematical models of viscosity - stratified flow in a hele - shaw cell .
these 1d models are suitable for description of the main features of the two - dimensional flow
. stationary solutions of eqs . obtained for the three - layer flow are in good agreement with the calculations of the flow on the basis of 2d model , ( see fig . [
fig : fig_6 ] ) . in the framework of the three - layer flow ( systems and ) the interpretation of the saffman
taylor instability is given .
solutions of eqs . and illustrating the formation of viscous fingers are constructed ( fig .
[ fig : fig_7 ] and [ fig : fig_8 ] ) . however , these models fail to predict the correct growth rate of the fingers .
we propose modification of the layered flow model in order to agree with this behaviour .
the model is obtained by including friction between the fluid layers .
this provides a non - zero thickness of the fingertip and under some assumptions allows one to describe the evolution of viscous fingers on the basis of the scalar equation with non - convex flux .
although the proposed equation involves empirical parameter , this model reveals the physical mechanism to ensure correct propagation velocity and structure of viscous fingers . as it can be seen from fig .
[ fig : fig_11 ] the velocity of propagation and the thickness of the fingers in proposed model ( if the empirical parameter @xmath253 is appropriately specified ) are in fairly good agreement with calculations based on two - dimensional equations , . comparison with the koval model ( see fig . [
fig : fig_10 ] ) confirms the correctness of the results .
this work was supported by the russian science foundation ( grant no .
15 - 11 - 20013 ) .
liapidevskii , v. yu . & teshukov , v. m. ( 2000 ) _ mathematical models of propagation of long waves in a non - homogeneous fluid .
_ novosibirsk .
siberian division of the russian academy of sciences .
( in russian ) .
smirnov , n. n. , nikitin , v. f. , maximenko , a. , thiercelin , m. & legros , j. c. ( 2005 ) instability and mixing flux in frontal displacement of viscous fluids from porous media .
fluids _ * 17 * , 084102 .
zimmerman , w. b. & homsy , g. m. ( 1992 ) viscous fingering in miscible displacements : unification of effects of viscosity contrast , anisotropic dispersion , and velocity dependence of dispersion on nonlinear finger propagation .
fluids a _ * 4 * , 23482359 . | a hierarchy of mathematical models describing viscosity - stratified flow in a hele - shaw cell is constructed .
numerical modelling of jet flow and development of viscous fingers with the influence of inertia and friction is carried out .
one - dimensional multi - layer flows are studied . in the framework of three - layer flow the interpretation of the saffman
taylor instability is given .
a kinematic - wave model of viscous fingering taking into account friction between the fluid layers is proposed .
comparison with calculations on the basis of two - dimensional equations shows that this model allows to determine the velocity of propagation and the thickness of the viscous fingers .
viscosity - stratified flow in a hele - shaw cell + a.a .
chesnokov@xmath0 , v.yu .
liapidevskii@xmath0 + @xmath1novosibirsk state university , + pirogova str .
2 , novosibirsk , 630090 , russia + @xmath2lavrentyev institute of hydrodynamics sb ras , + lavrentyev ave . 15 , novosibirsk , 630090 , russia + e - mails : [email protected] , [email protected] keywords : hele - shaw flow , fingering instability , kinematic - wave model . |
there is a well - known positive correlation between galaxy luminosity and metallicity @xcite .
@xcite measured this relation for 53400 star - forming galaxies from the sloan digital sky survey ( sdss ) data release 4 ( dr4 , @xcite ) , estimating the gas phase oxygen abundance , @xmath3 , from region emission lines as the surrogate for `` metallicity . '' using estimates of the galaxy stellar masses derived from the techniques of @xcite , @xcite showed that metallicity is better correlated with galaxy mass than with luminosity , which suggests that the observed relation arises from the fact that lower mass galaxies have lower escape velocities than higher mass galaxies and so lose metals more easily via , e.g. , supernova winds @xcite .
however , @xcite showed that even if a low - mass galaxy can selectively remove its metals via winds , subsequent star formation can essentially erase the effects on the gas - phase metallicity ; both a low star formation rate and a large gas fraction are needed to retain low metal abundances .
other proposed origins for the mass metallicity relation include lower star formation rates in less massive galaxies giving rise to fewer massive stars and therefore less metal production @xcite . despite the locus of galaxies in the mass metallicity plane having intrinsically high scatter ( e.g. , a 1-@xmath4 spread of @xmath5 dex in @xmath3 at @xmath6 ) , in the absence of good quality spectra
it has become an increasingly common practice to deduce or assign metallicities to galaxies based simply on their luminosities . in particular , the idea persists that dwarf galaxies are always `` metal - poor , '' certainly a fair assumption in the local group . in this study however , we will show that there is a significant population of low - mass galaxies with high gas - phase oxygen abundances relative to the mass metallicity relation defined by @xcite ; high - mass low - metallicity outliers from the mass metallicity relation will be the focus of an upcoming paper . as shown in figure [ fig : ohmb ]
, we find 24 galaxies with @xmath7 mag and @xmath8 and 17 high - metallicity outliers with @xmath9 mag .
we explain how we selected this sample in [ sec : outliers ] and verified the galaxies high oxygen abundances in [ sec : metal ] . in [ sec : disc ] we discuss some of the possible explanations for the existence of such a population of galaxies and explain why we predict that these galaxies should have relatively low gas masses .
specifically , in [ sec : trans ] we discuss the evidence for our galaxies being so - called `` transition '' dwarf galaxies .
finally , we summarize our results and conclusions in [ sec : conc ] .
we began with the sample of @xmath10 star - forming galaxies with measured gas - phase oxygen abundances and stellar masses from @xcite .
high - metallicity outliers from the mass metallicity locus can be due to main three causes , and our cuts were taken with these in mind .
first , some galaxies can spuriously appear to be underluminous for their mass .
examples include highly inclined galaxies subject to strong internal extinction and regions in larger galaxies that sdss has mistakenly flagged as a galaxy .
among relatively nearby galaxies ( e.g. , those in the virgo cluster ) some have large peculiar velocities relative to the hubble flow , leading to a misestimation of their distance modulus and thus absolute magnitude .
secondly , rogue outliers can have spuriously high estimated metallicities ; the galaxy , for example , may not be at high enough redshift to have the strongly constraining [ ] @xmath11 emission line in the sdss bandpass .
also , the strengths of the [ ] and [ ] lines paradoxically _ decrease _ with oxygen abundance , so at low signal - to - noise , the metallicity can be overestimated due to underestimated line fluxes . finally , the object can be an honest outlier ; these are the objects we want in our final sample .
we found that a large number of cuts in different parameter spaces is useful for automatically rejecting many objects which would otherwise have to be thrown out by visual inspection .
we ran two searches on the full @xcite sample for high - metallicity outliers .
for the first , we were very conservative with our cuts so as to be certain that the remaining galaxies are both statistically significant and not spurious .
however , due to significant contamination at low luminosities , this `` main '' sample has no members with @xmath12 .
we therefore did a separate , less stringent search for the lowest luminosity outliers ; to distinguish it from the main sample , we refer to this sample of lower mass galaxies as the `` very low mass '' sample .
figure [ fig : ohmb ] shows where our sample falls in the @xmath3@xmath13 plane ; all galaxy images are shown in figure [ fig : images ] and summary information is presented in table [ tbl : sample ] . for reference , in figure
[ fig : ohmb ] we plot the small and large magellanic clouds and milky way using @xmath13 from @xcite .
we recalculated @xmath3 for the smc and lmc using the line ratios reported by @xcite of six regions in the smc and four regions in the lmc and the relation @xmath14 , where @xmath15\lambda6584/\mbox{h}\alpha)$ ] @xcite .
we then transformed these abundances onto the @xcite scale using the formulae given by @xcite , resulting in average metallicities of 8.21 and 8.39 for the smc and lmc , respectively .
the milky way is plotted also plotted for reference at the solar oxygen abundance of 8.86 @xcite .
, the solar abundance is neither a nebular abundance nor the abundance within the central @xmath16kpc of the galaxy . ]
rrclcrl 193.9056 & @xmath17 & 8.68 & @xmath18 & 7.65 & 0.0029 & blue core , ic 225 + 180.4589 & @xmath19 & 8.73 & @xmath20 & 7.73 & 0.0035 & + 190.2097 & @xmath21 & 9.06 & @xmath22 & 7.39 & 0.0025 & blue core , vcc 1855 + 179.0295 & @xmath23 & 8.83 & @xmath24 & 8.06 & 0.0047 & + 228.0340 & @xmath25 & 8.76 & @xmath26 & 8.21 & 0.0065 & + 126.6407 & @xmath27 & 8.69 & @xmath28 & 8.07 & 0.0072 & + 227.2679 & @xmath29 & 8.85 & @xmath30 & 8.17 & 0.0055 & + 193.6735 & @xmath31 & 9.12 & @xmath32 & 8.40 & 0.0029 & bright core + 126.6633 & @xmath33 & 8.86 & @xmath34 & 8.59 & 0.0078 & + 208.3607 & @xmath35 & 8.94 & @xmath36 & 8.64 & 0.0027 & blue core + 128.5841 & @xmath37 & 8.95 & @xmath38 & 8.69 & 0.0114 & + 40.3408 & @xmath39 & 8.90 & @xmath40 & 8.70 & 0.0227 & + 212.9768 & @xmath41 & 8.85 & @xmath42 & 8.81 & 0.0064 & blue core + 190.5886 & @xmath43 & 9.00 & @xmath44 & 8.40 & 0.0044 & bright core + 36.6179 & @xmath45 & 8.82 & @xmath46 & 7.92 & 0.0051 & blue core + 117.1775 & @xmath47 & 8.94 & @xmath48 & 9.08 & 0.0155 & + 182.7831 & @xmath49 & 8.95 & @xmath50 & 9.13 & 0.0206 & + 133.8883 & @xmath51 & 9.10 & @xmath52 & 9.29 & 0.0068 & + 25.7673 & @xmath53 & 9.12 & @xmath54 & 9.25 & 0.0284 & + 225.5670 & @xmath55 & 9.12 & @xmath56 & 9.45 & 0.0147 & bright core + 142.9797 & @xmath57 & 9.07 & @xmath58 & 9.35 & 0.0275 & + 139.3797 & @xmath59 & 9.15 & @xmath60 & 9.46 & 0.0221 & + 202.2088 & @xmath61 & 9.14 & @xmath62 & 9.26 & 0.0217 & + 39.0487 & @xmath63 & 9.12 & @xmath64 & 9.16 & 0.0314 & + 143.4695 & @xmath65 & 9.14 & @xmath66 & 9.54 & 0.0145 & bright core + 187.1760 & @xmath67 & 9.13 & @xmath68 & 9.35 & 0.0239 & + 182.8574 & @xmath69 & 9.14 & @xmath70 & 9.52 & 0.0231 & + 211.3199 & @xmath71 & 9.13 & @xmath72 & 9.54 & 0.0417 & + 258.4301 & @xmath73 & 9.16 & @xmath74 & 9.61 & 0.0290 & + 208.9341 & @xmath75 & 9.14 & @xmath76 & 9.68 & 0.0295 & + 227.6711 & @xmath77 & 9.13 & @xmath78 & 9.61 & 0.0316 & + 352.8635 & @xmath79 & 9.14 & @xmath80 & 9.74 & 0.0324 & + 206.4402 & @xmath81 & 9.14 & @xmath82 & 9.58 & 0.0313 & + 153.2177 & @xmath83 & 9.20 & @xmath84 & 9.57 & 0.0315 & + 177.7870 & @xmath85 & 9.14 & @xmath86 & 9.41 & 0.0482 & + 144.1848 & @xmath87 & 9.14 & @xmath88 & 9.71 & 0.0426 & + 162.6325 & @xmath89 & 9.16 & @xmath90 & 9.68 & 0.0384 & + 195.6390 & @xmath91 & 9.26 & @xmath92 & 9.55 & 0.0471 & + 235.5660 & @xmath93 & 9.16 & @xmath94 & 9.83 & 0.0425 & + 163.2191 & @xmath95 & 9.27 & @xmath96 & 9.90 & 0.0242 & + 245.3673 & @xmath97 & 9.25 & @xmath98 & 9.91 & 0.0285 & blue core + table [ tbl : cuts ] summarizes our selection of the main sample of 24 metal - rich low - mass galaxies .
spiral galaxies are known to have radial metallicity gradients such that the nuclear abundances are higher than averages over whole galaxies .
thus , if only a small fraction of the galaxy is within the sdss spectroscopic @xmath99 diameter fibers , the measured metallicity can appear to be artificially high relative to other galaxies .
following @xcite , we make an initial cut by requiring that the fraction of the galaxy covered by the fiber to be greater than 10% .
( in our final cut , we follow @xcite and change this lower bound to 20% , which only eliminates two of the galaxies contained in the penultimate sample listed in table [ tbl : cuts ] . )
we wanted the sample to be statistically significant , so outliers were then determined via a series of cuts in the @xmath3 versus absolute @xmath100-band magnitude ( @xmath13 ) , @xmath101-band ( @xmath102 ) , and stellar mass ( m@xmath103 ) planes , as demonstrated with m@xmath103 in figure [ fig : cuts ] . for example , we divided the 52477 objects with sdss magnitude errors @xmath104 mag in all bands into bins of @xmath13 of width @xmath105 mag ; in each bin we kept the 2.5% with the highest @xmath3 .
we likewise took bins of @xmath3 of width 0.1 dex and kept 2.5% of the objects with the faintest @xmath13 .
similar cuts were made with @xmath102 ( binsize @xmath106 mag ) and @xmath107 ( binsize of @xmath108 dex ) .
@xmath102 was calculated from the sdss @xmath101-band magnitude and the spectroscopic redshift , @xmath13 was calculated using the low - resolution spectral templates of @xcite , and the stellar masses from the @xcite sample were measured using a combination of sdss colors and spectra as described by @xcite .
only 58 objects survived this series of cuts . at this point
, we made a redshift cut : galaxies must have @xmath109 in order to have the [ ] @xmath110 emission line pair in their sdss spectrum , and as this is a highly constraining line for the metallicity , we determined that it should be in the spectrum in order to remove a possible source of systematics and so that we can measure the metallicity with a diagnostic that uses this line .
seven @xmath111 galaxies pass our subsequent @xmath3 error and visual inspection cuts ; after studying their spectra , as discussed below , we chose to keep these galaxies in our main sample .
we also forced the cited @xmath112 error in @xmath3 to be less than 0.05 dex ; metallicities with large uncertainties are obviously more likely to be spurious than ones with small uncertainties .
unsurprisingly , the @xmath113 redshift cut ( after all of the other cuts ) also removed any galaxies with fiber fractions @xmath114 which would have survived to that stage .
we find , however , that this large number of cuts in various parameter spaces dramatically reduced the number of objects that had to be removed `` by eye . ''
the final visual - inspection cut removed only two objects with nearby potential photometric contaminations .
finally , we re - examined the fiber fraction cut ; following @xcite we allowed the lower - limit fiber fraction to be 20% ( excluding two objects from the sample ) .
the final exclusion was a barred galaxy with a diameter of @xmath115 ; the sdss fiber size is @xmath99 , but this galaxy is labeled as having a fiber fraction of @xmath116% .
this galaxy probably has spuriously low sdss petrosian magnitudes .
the 24 galaxies in this main sample have a range of fiber fractions between 25% and 60% and a redshift range of @xmath117 .
ll fiber fraction @xmath118 & 107992 + @xmath119-corrected sdss magnitude errors @xmath104 mag & 48327 + 97.5% large ( o / h ) and small @xmath13 & 227 + 97.5% large ( o / h ) and small @xmath102 & 202 + 99% large ( o / h ) w.r.t .
@xmath13 & 87 + 99% large ( o / h ) w.r.t .
m@xmath103 & 66 + 97.5% small m@xmath103 w.r.t . ( o / h ) & 51 + redshift @xmath120 for [ ] & 34 + @xmath3 error @xmath121 dex & 22 + visual inspection & 20 + fiber fraction @xmath122 and photometry & 17 + redshift @xmath123 but surviving subsequent cuts & @xmath124 it is obvious from figure [ fig : ohmb ] that the selection of our main sample artificially imposes a lower mass cutoff .
this does not mean that there is not a statistically significant sample of interesting galaxies with @xmath9 ; it just means that there are more contaminating objects with measured high metallicities in the low - luminosity regime than at brighter magnitudes .
that is , the most extreme outliers with @xmath125 in one parameter space ( e.g. , the @xmath3 vs. @xmath13 plane ) are not also outliers in one of the others ( e.g. , the @xmath3 vs. @xmath107 plane ) .
the _ true _ very low luminosity mass metallicity outliers therefore have lower abundances than these spurious outliers .
we therefore did a separate search for high - metallicity galaxies at extremely low luminosities and masses , as summarized in table [ tbl : lowmcuts ] .
we chose to not apply a fiber fraction cut for this sample because lower luminosity galaxies are preferentially closer and therefore subtend a larger angle on the sky , making it more difficult for a substantial fraction of the galaxy to be within the 3 sdss fiber diameter .
we therefore began the selection with a magnitude error cut , like the one for the main sample .
we also removed objects with fiber fractions @xmath126 occupying the parameter space already excluded by the main sample : any objects with @xmath13 brighter than @xmath127 mag , @xmath128 , and a high @xmath129 at the 95% level with respect to m@xmath103 were excluded .
the cuts in @xmath3 relative to @xmath13 , @xmath102 and m@xmath103 are all less stringent than for the main sample , but because for this sample we are interested in the _ very _ low mass objects , we took a strong ( 99% level ) cut in m@xmath103 relative to @xmath3 .
six galaxies were excluded due to metallicity re - estimation considerations ( as discussed below ) .
three of the remaining galaxies are potentially members of the virgo cluster ; of these , only two remain clear outliers from the @xmath13metallicity locus when their distance moduli are shifted to the virgo value of @xmath130 magnitudes @xcite .
( we do not attempt to correct the stellar mass estimates due to the effects of peculiar velocity . ) finally , the 95% high ( o / h ) with respect to @xmath102 cut excludes the galaxy ic 225 from the sample , which is pointed out by @xcite to have a relatively high metallicity and a compact blue core .
as this galaxy passes all of the other cuts and is clearly interesting , we added it back to the sample , raising the total number of galaxies in the very low mass sample to 17 and the full sample to 41 .
the very low mass sample has fiber fractions ranging from 0.04 to 0.45 ; more than half of the 17 galaxies in this sample have fiber fractions below 0.1 and only 5 are above 0.2 .
also , none of these galaxies would have survived the main sample redshift cut , as the very low mass galaxies occupy the redshift range @xmath131 , with the most nearby objects being only @xmath132 mpc away .
ll @xmath119-corrected sdss magnitude errors @xmath104 mag & 52744 + 84% high ( o / h ) w.r.t . @xmath13
& 8405 + @xmath9 mag & 217 + @xmath133 & 201 + 84% high ( o / h ) w.r.t .
m@xmath103 & 192 + @xmath3 error @xmath121 dex & 139 + 99% low @xmath107 w.r.t .
( o / h ) & 105 + _ exclude _ 95% high ( o / h ) w.r.t .
m@xmath103 , fiber fraction @xmath134 & 78 + 95% high ( o / h ) w.r.t .
@xmath102 & 32 + visual inspection & 23 + metallicity comparison & 17 + absolute magnitude correction & 16 are these 41 galaxies true outliers from the mass metallicity relation , or are they just the tail of the scatter of the @xcite measurements ? as a first check , the galaxies fall where expected on the standard @xcite ( bpt ) diagrams for @xmath135}\,\lambda 5007/\mbox{h}\beta)$ ] vs.@xmath136}\,\lambda 6548/\mbox{h}\alpha)$ ] and @xmath137}\,\lambda\lambda 6717 + 31)/\mbox{h}\alpha)$ ] @xcite .
we note that there are several complications with measuring high metallicities using visible wavelength spectra ( see @xcite for a thorough review ) .
the main problem is that cooling is more efficient at high metallicities , which translates into lower nebular temperatures and hence weaker [ ] and [ ] visible - wavelength lines , with the primary cooling load shifting to the far - infrared fine structure emission lines that can not be readily observed .
the result is that at high abundances the visible wavelength spectra become increasingly insensitive to changes in @xmath129 .
in fact , some diagnostics , like the traditional r@xmath138 line index , effectively saturate at high metallicity @xcite , making it rather difficult to accurately measure even relative abundances . because we would like to verify the high estimated oxygen abundances and we can not reproduce the @xcite abundance calculations , which are based on a bayesian statistical method , we measured @xmath129 in our galaxies using two recommended methods from @xcite .
another possible source of systematic inaccuracy in the abundance estimates is the treatment of extinction corrections .
the corrections traditionally applied use the balmer decrement and assume a simple uniform foreground obscuring screen model like that used for stars to estimate @xmath139 and correct the other emission lines .
it is expected , however , that extinction towards extended sources like regions is better described as a clumpy screen , for example the turbulent screen models of @xcite . in these , use of a simple screen tends to systematically _ overestimate _ the extinction correction for the @xmath1403727,29 emission line , leading to a systematic _ underestimate _ of the gas - phase oxygen abundance . in all of our galaxies
the balmer decrement measurements are consistent with low @xmath139 ( @xmath141 ) for a simple screen extinction model , so this is not a big effect compared to other sources of measurement error given the attenuation curves in @xcite .
@xcite have measured the gas - phase metallicities in star forming galaxies from sdss using ten different methods , including that of @xcite ; they also provide average relations relating each pair of methods . while no one metallicity estimate is strictly believable
, the _ true _ oxygen - to - hydrogen ratio the _ relative _ measurements are generally robust @xcite . using the data release 6 sdss spectra @xcite
, we subtracted the underlying stellar continuum using the starlight program @xcite .
we then calculated @xmath3 using the revised @xcite method given in equation ( a3 ) of @xcite , @xmath142}/\mbox{[\ion{o}{2 } ] } ) & = & 1106.8660 - 532.1451z + 96.37326z^2 \\ & & - 7.8106123z^3 + 0.32928247z^4,\end{aligned}\ ] ] where @xmath143 .
like @xcite , we found the roots of this equation using the ` fz_roots ` program in idl .
we also measured the metallicity using the `` o3n2 '' method of @xcite , as recommended by @xcite , where @xmath144\lambda 5007/\mbox{h}\beta}{\mbox{[\ion{n}{2}]}\,\lambda 6584/\mbox{h}\alpha } \right).\ ] ] we choose to not use the r@xmath138 diagnostic for measuring @xmath3because it is known to be difficult to calibrate at these abundances @xcite .
in particular , while the seventeen @xmath109 galaxies in our main sample span @xmath145 dex on the @xcite scale , they span @xmath146 dex when their metallicities are calculated using the r@xmath138 methods of either @xcite or @xcite , reflecting the fact that the r@xmath138 parameter essentially saturates at high oxygen abundances @xcite .
the oxygen abundances for our 41 galaxies plotted against these adopted metallicities in figure [ fig : convz ] .
the dotted line indicates equal measurements ; the other two lines are the empirically determined relations from @xcite .
the galaxies in our sample fall preferentially above the relation for the @xcite method , but surprisingly those galaxies for which it is measurable fall _ below _ the relation for the @xcite method .
one possible explanation for this discrepancy is that because [ ] and [ ] are widely separated in wavelength , a misestimation of the extinction or a poorly fitted continuum ( especially for the weak [ ] emission line ) can lead to underestimated abundances .
specifically , there is often a degeneracy when fitting the stellar continuum between the extinction and the stellar population .
while the continuum fitting for our spectra are fairly good ( i.e. , they exhibit low residuals after the stellar template is subtracted ) , we tested the sensitivity to the continuum level at [ ] @xmath147by calculating the @xcite metallicity when the continuum is over- and underestimated ; the mean of these two extreme metallicities is still systematically larger than expected by the mean relation , implying that the shift is not due to continuum fitting error .
we quantify the differences between the @xcite and @xcite methods by examining the product of the vertical displacement of each galaxy ( for an optimally fit continuum ) from the two @xcite relations ; only one galaxy in the main sample has a significantly high @xcite metallicity relative to both of the other indicators .
however , this @xmath3 was still high enough that even when the @xcite metallicity was replaced with the one expected from either relation , the galaxy still passes the cuts in both the @xmath102- and m@xmath103-metallicity planes . while the @xmath148 galaxies do not have measurable [ ] ( and hence we can not use the @xcite diagnostic for them ) , the low redshift sample occupies a similar part of the @xcite versus @xcite parameter space as the galaxies in the main sample . for the very low - mass sample of galaxies , out of the 22 galaxies which passed our visual inspection test , we excluded 6 because they were more than @xmath149 dex above the @xcite relation from @xcite ; this is actually a more stringent cut than used for the main sample because several of the galaxies in the main sample with @xmath134 dex deviations from the @xcite relation have compensating negative deviations relative to the @xcite scale .
we interpret these results to mean that the 41 galaxies we have identified are true high - metallicity low - mass outliers from the mass metallicity relation .
the 41 dwarf galaxies selected as described in [ sec : outliers ] are surprisingly non - pathological .
they have undisturbed stellar morphologies ( see figure [ fig : images ] ) .
furthermore , given that by selection these galaxies have both low luminosities and high metallicities and that there is a trend for redder , brighter galaxies to have higher oxygen abundances ( see , e.g. , @xcite ) these metal - rich dwarfs do not occupy unexpected region of the color magnitude diagram ( see figure [ fig : cmd ] and also [ sec : trans ] ) .
how , then , did they come to have such high oxygen abundances ?
one obvious possibility is that these metallicities are due to an environmental effect , but as we describe in [ sec : enviro ] , the galaxies in our sample are non - interacting and rather isolated . though several models predict that these galaxies should have high specific star formation rates , we show in [ sec : sfr ] , that these galaxies have normal star formation rates for their masses .
we explain in [ sec : gasfrac ] why we predict that these metal - rich dwarf galaxies should have relatively low gas fractions , and we discuss the implications of this prediction in the broader context of transition - type dwarf galaxies in [ sec : trans ] .
while the origin of scatter in the mass metallicity relation is unknown , one popular proposal is that environment affects metallicity .
recently , @xcite studied close pairs of star - forming galaxies in sdss with projected separations of @xmath150kpc and radial velocity separations of @xmath151kms@xmath152 .
their results indicate that for minor mergers , the less massive galaxy is likely to be preferentially more metal rich than predicted by the mass metallicity relation .
likewise , @xcite , after accounting for correlations within the color magnitude diagram , find a strong positive correlation between metallicity and overdensity for @xmath153 sdss - selected star - forming galaxies that can account for up to @xmath154% of the scatter in the mass metallicity relation . in light of these results , we searched for nearby neighbors of the seventeen @xmath109 main sample metal - rich dwarf galaxies in a cylindrical volume of depth @xmath155kms@xmath152 and projected radius of 1 mpc .
we find that our galaxies are relatively isolated ; none of them would have made it into the @xcite sample .
seven out of these seventeen galaxies in our sample have no neighbors within this volume . of the 10 remaining galaxies , four have no neighbors within 500kms@xmath152 and 500kpc , and only two have any neighbors within 500kms@xmath152 and 100kpc .
one of these two seems to be on the outskirts of a nearby cluster , but it is unclear whether or not it is physically associated with the cluster ( and there is only one small , faint , non - star - forming galaxy within the @xcite volume ) .
the other galaxy with a nearby neighbor is @xmath156kms@xmath152 and 85kpc from a less - massive relatively metal - poor ( @xmath157 \approx 8.6 $ ] ) star - forming galaxy .
we therefore conclude that interactions with neighbors do not explain the observed high abundances of our main sample of @xmath158 galaxies .
the galaxies in the very low mass sample ( as well the @xmath123 galaxies in the main sample ) , however , are generally at low enough redshift that an automated search for neighbors in velocity and projected distance space is difficult . like the galaxies in the main sample , none of the lower - mass galaxies are in obviously interacting systems .
only one has an clear companion ; it is a @xmath159 galaxy which appears to be a satellite of a non - starforming companion 33kpc and @xmath160kms@xmath152 away .
also , while two of the seventeen galaxies in the very low mass sample may be in the virgo cluster , clearly a rich environment can not explain the high oxygen abundances observed in all of these galaxies .
@xcite finds that for a galaxy to have a low metallicity , it must have both a low star formation rate and a high gas fraction ; she argues that the observed low star formation efficiency in galaxies with circular velocities @xmath161kms@xmath152 can explain the mass metallicity relation .
@xcite similarly suggest that the mass metallicity relation could be due to a mass
star formation rate relation : a galaxy with a lower rate of star formation will have relatively fewer massive stars , and therefore a lower oxygen abundance . in either of these scenarios
, we might expect the mass metallicity outliers to have high star formation rates for their masses .
on the other hand , @xcite find that at low stellar masses , galaxies with higher specific star formation rates tend to have _ lower _ oxygen abundances . as figure [ fig : sfrm ] shows , the galaxies in the main sample do not have preferentially high or low instantaneous specific star formation rates , while typical star formation rates for the the very low mass sample galaxies are @xmath162dex _ lower _ than expected given their masses .
so why do these galaxies have such high oxygen abundances ?
consider a closed box star forming system ( i.e. , a system with no gas inflow or outflow ) .
the metallicity @xmath163(mass of metals in gas phase)/(total gas mass ) is @xmath164 where @xmath165 is the gas fraction and @xmath166(mass of metals in gas)/(mass of metals in stellar remnants and main sequence stars ) is the metal yield .
an immediately striking aspect of equation [ eqn : closedbox ] is that there is no implicit dependence on the total galaxy mass ; the deviations from a universal metallicity in equation [ eqn : closedbox ] is _ not _ the same as @xmath3it is a mass ratio of _ all _ metals rather than the abundance ratio of one element relative to hydrogen the same arguments still qualitatively hold for observed abundances .
] observed via the mass metallicity relation are presumably then due to the fact that galaxies are not scaled closed box versions of one another : variations of the yield with mass , a dependence on the gas fraction with mass , or a combination of these effects plays a role . for a galaxy to have a higher metallicity than other galaxies of the same mass , equation [ eqn : closedbox ]
tells us that it must have either a relatively high yield or a relatively low gas fraction .
observationally , the yield which depends on both star formation physics as well as gas inflow and outflow is a difficult quantity to measure ; one popular way to address this problem is to define an effective yield , @xmath167,\ ] ] as the yield the galaxy _ would _ have were it actually a closed system .
given the galaxy s metallicity and gas fraction , one can then easily calculate @xmath168 .
a common explanation for the mass metallicity relation is that the effective yield is positively correlated with galaxy mass , so that lower mass galaxies are able to preferentially lose metals to the intergalactic medium via winds because of their relatively shallow potential wells ( see e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) .
in particular , @xcite tested this idea by comparing effective yields and baryonic masses .
however , because when estimating the effective yield from equation [ eqn : yeff ] , @xcite use the measured @xmath3 values , outliers in the mass metallicity relation are practically guaranteed to be outliers in the effective yield
baryonic mass relation .
regardless , it is obvious from equation [ eqn : closedbox ] that our metal - rich dwarfs must either have unusually high yields or unusually low gas fractions for their masses .
galaxy yields can be affected by three processes : star formation , metal - deficient gas inflows , and metal - rich gas outflows .
as shown in figure [ fig : sfrm ] and discussed in [ sec : sfr ] , we find that the star formation rates for this population of galaxies are consistent with those of other galaxies of similar masses .
@xcite has shown that metal - poor gas inflow is insufficient to explain the typically low metal abundances of low - mass galaxies ; we therefore conclude that a lower inflow rate is insufficient to explain the higher abundances of our galaxies .
it is possible that these galaxies are less effective at driving outflows than other dwarfs .
for example , @xcite find that , at fixed mass , galaxies with smaller half - light radii tend to have higher abundances ; this picture is consistent with the idea that it is more difficult to drive winds from deeper potential wells .
( @xcite , on the other hand , find no correlation with how concentrated a galaxy s light is and its oxygen abundance . ) while the galaxies in our main sample do tend to have small radii for their masses ( half - light radii of less than 2kpc ) , the typical @xmath3 for galaxies of similar masses and radii is still @xmath162dex lower ( about @xmath169 ) than that of our galaxies . furthermore , the very low mass sample galaxies do not have preferentially small radii for their masses , leading us to conclude that while small radii may be a contributing factor to why some of these galaxies have been able to retain their metals , size alone does not tell the whole story .
@xcite calculated that enriched gas outflows can only severely decrease a galaxy s effective yield if the gas fraction is sufficiently high ; that is , for low gas fractions , even a very strong outflow can not drastically decrease the effective yield and thus measured abundance .
likewise , if a galaxy has a relatively low gas fraction , then only a small amount of pollution is needed to enrich the gas and cause the measured abundance to be high .
we therefore predict that these mass metallicity outliers have anomalously low gas masses relative to other isolated galaxies of similar luminosities and star - formation rates .
there are unfortunately no data in the literature for our galaxies that we can call upon to lend observational support to this prediction .
however , @xcite have found that , in a sample of 27 nearby dwarf irregular galaxies , the gas - phase oxygen abundance is negatively correlated with the -measured gas - to - stellar mass ratio , which lends observational credence to our expectation .
if these high - metallicity dwarf galaxies really do have relatively little gas , then they should be rapidly approaching the end of their star formation .
specifically , these galaxies are likely to be transitioning from gas - rich dwarf irregulars ( dirr ) to gas - deficient dwarf spheroidals ( dsph ) or the more massive dwarf ellipticals ( de ) . in general , so - called dirr / dsph transitional dwarfs have similar star - formation histories as their currently non - starforming dsph cousins : both galaxy types typically have a mix of old and intermediate - age stellar populations @xcite . in a study of five nearby star - forming transitional dwarfs
, @xcite found that these relatively isolated galaxies seem to have unusually high oxygen abundances for their luminosities , much like our sample of mass metallicity outliers . ) , highlighting the point that while our sample is rather pure , it is almost certainly incomplete . ]
@xcite stress that the only difference between dsph galaxies and the transitional dirr / dsph galaxies is the absence of star formation and of gas in dsph galaxies .
while much of this conclusion is based on galaxies in higher - density environments than ours are ( i.e. , the @xcite dwarf galaxies are mostly in the local group ) , we find it likely that our galaxies are part of a similar transition population : morphologically , they have smooth , undisturbed profiles ( see figure [ fig : images ] ) , indicating a stronger relationship to dwarf spheroidals than to the dwarf irregulars .
in particular , for all of our galaxies we note a lack of the irregular flocculent structures or spiral features often associated with dirr galaxies .
in fact , the only remarkable morphology any of the galaxies in our sample displays are the bright and often very blue cores noticed in ten of the galaxies .
these galaxies are shown in figure [ fig : blue ] and noted in table [ tbl : sample ] . some of these cores appear to have merely much higher surface brightnesses than their surroundings , but some are decidedly blue : @xcite measure a difference of @xmath170 mag arcsec@xmath171 from the center of ic 225 , ic 255 did not pass all of the cuts for our very low mass sample , and so we reinserted it into the sample by hand . ]
relative to isophotes at @xmath172 .
these bright cores are almost certainly not a signature of active galactic nuclei ; as mentioned in [ sec : metal ] , these galaxies fall squarely within the star - forming locus of the normal bpt line - ratio diagnostic diagrams @xcite .
@xcite find that for similar transitional dwarf galaxies with blue centers in the virgo cluster ( including vcc 1855 , the least massive galaxy in our sample with @xmath173\sim 7.4 $ ] ) , star formation is limited to these central blue nuclei .
( @xcite attribute this centralized star formation to the effects of ram pressure stripping preferentially removing gas from the outer regions of the galaxies . however , not all of the galaxies in our sample with bright nuclei are in the kinds of rich environments like the virgo cluster where ram pressure stripping expected to play an important role . ) intriguingly ( yet unsurprisingly ) , most of our galaxies which appear to be approaching or on the red sequence ( see figure [ fig : cmd ] ) also have bright and/or blue cores , implying that their colors are already dominated by their non - starforming regions . furthermore
, having star formation limited to a relatively small region of the galaxy supports the idea that there is relatively little fuel available for forming stars , and thus that these galaxies are on their way to becoming standard dwarf spheroidal or dwarf elliptical galaxies .
we have identified a sample of 41 low - luminosity high oxygen abundance outliers from the mass metallicity relation . 1 .
these galaxies are fairly isolated and none show any signs of morphological disturbance .
it is likely , however , that the handful of galaxies in this sample which are possible group or cluster members have disturbed _ gas _ morphologies , due to e.g. , ram - pressure stripping .
however , environmental effects can not account for the large gas - phase oxygen abundances observed in all of these galaxies .
2 . based on the effective yield considerations of @xcite and various related observational results @xcite
, we predict that these galaxies should have low gas fractions relative to other dwarf galaxies of similar stellar masses and star formation rates in similar environments but with more typical ( i.e. , lower ) oxygen abundances .
furthermore , in this scenario , the _ stellar _ metallicities should be consistent with that of other dwarfs of similar luminosity , i.e. , the stellar metallicities should be lower than the measured gas - phase abundances .
3 . these mass metallicity outliers appear to be `` transitional '' dwarf galaxies : because they are running out of star - forming fuel , they are nearing the end of their star formation and becoming typical isolated dwarf spheroidals .
this conclusion is most sound for the @xmath174 galaxies which are observed to have low specific star formation rates .
more data specifically , gas fractions and stellar metallicities are needed to help elucidate the situation for the higher mass galaxies .
we would like to thank roberto assef for adapting his low - resolution template code for calculating @xmath13 for the sdss sample for us , cayman unterborn for letting us use his data and code for generating the color magnitude diagram in figure [ fig : cmd ] , and sara ellison for giving us access to some of the data from @xcite .
we are grateful to the anonymous referee , david weinberg , christy tremonti , and thorsten lisker for helpful comments on the text .
we would also like to thank jos luis prieto , jeff newman , mike cooper , kristian finlator , john moustakas , henry lee , paul martini , and todd thompson for useful discussions and suggestions .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is ` http://www.sdss.org/ ` .
the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington . , j. k. , ageros , m. a. , allam , s. s. , allende prieto , c. , anderson , k. s. j. , anderson , s. f. , annis , j. , bahcall , n. a. , bailer - jones , c. a. l. , baldry , i. k. , barentine , j. c. , bassett , b. a. , becker , a. c. , beers , t. c. , bell , e. f. , berlind , a. a. , bernardi , m. , blanton , m. r. , bochanski , j. j. , boroski , w. n. , brinchmann , j. , brinkmann , j. , brunner , r. j. , budavri , t. , carliles , s. , carr , m. a. , castander , f. j. , cinabro , d. , cool , r. j. , covey , k. r. , csabai , i. , cunha , c. e. , davenport , j. r. a. , dilday , b. , doi , m. , eisenstein , d. j. , evans , m. l. , fan , x. , finkbeiner , d. p. , friedman , s. d. , frieman , j. a. , fukugita , m. , gnsicke , b. t. , gates , e. , gillespie , b. , glazebrook , k. , gray , j. , grebel , e. k. , gunn , j. e. , gurbani , v. k. , hall , p. b. , harding , p. , harvanek , m. , hawley , s. l. , hayes , j. , heckman , t. m. , hendry , j. s. , hindsley , r. b. , hirata , c. m. , hogan , c. j. , hogg , d. w. , hyde , j. b. , ichikawa , s .-
i . , ivezi , . ,
jester , s. , johnson , j. a. , jorgensen , a. m. , juri , m. , kent , s. m. , kessler , r. , kleinman , s. j. , knapp , g. r. , kron , r. g. , krzesinski , j. , kuropatkin , n. , lamb , d. q. , lampeitl , h. , lebedeva , s. , lee , y. s. , leger , r. f. , lpine , s. , lima , m. , lin , h. , long , d. c. , loomis , c. p. , loveday , j. , lupton , r. h. , malanushenko , o. , malanushenko , v. , mandelbaum , r. , margon , b. , marriner , j. p. , martnez - delgado , d. , matsubara , t. , mcgehee , p. m. , mckay , t. a. , meiksin , a. , morrison , h. l. , munn , j. a. , nakajima , r. , neilsen , jr .
, e. h. , newberg , h. j. , nichol , r. c. , nicinski , t. , nieto - santisteban , m. , nitta , a. , okamura , s. , owen , r. , oyaizu , h. , padmanabhan , n. , pan , k. , park , c. , peoples , j. j. , pier , j. r. , pope , a. c. , purger , n. , raddick , m. j. , re fiorentin , p. , richards , g. t. , richmond , m. w. , riess , a. g. , rix , h .- w . , rockosi , c. m. , sako , m. , schlegel , d. j. , schneider , d. p. , schreiber , m. r. , schwope , a. d. , seljak , u. , sesar , b. , sheldon , e. , shimasaku , k. , sivarani , t. , smith , j. a. , snedden , s. a. , steinmetz , m. , strauss , m. a. , subbarao , m. , suto , y. , szalay , a. s. , szapudi , i. , szkody , p. , tegmark , m. , thakar , a. r. , tremonti , c. a. , tucker , d. l. , uomoto , a. , vanden berk , d. e. , vandenberg , j. , vidrih , s. , vogeley , m. s. , voges , w. , vogt , n. p. , wadadekar , y. , weinberg , d. h. , west , a. a. , white , s. d. m. , wilhite , b. c. , yanny , b. , yocum , d. r. , york , d. g. , zehavi , i. , & zucker , d. b. 2008 , , 175 , 297 , j. k. , ageros , m. a. , allam , s. s. , anderson , k. s. j. , anderson , s. f. , annis , j. , bahcall , n. a. , baldry , i. k. , barentine , j. c. , berlind , a. , bernardi , m. , blanton , m. r. , boroski , w. n. , brewington , h. j. , brinchmann , j. , brinkmann , j. , brunner , r. j. , budavri , t. , carey , l. n. , carr , m. a. , castander , f. j. , connolly , a. j. , csabai , i. , czarapata , p. c. , dalcanton , j. j. , doi , m. , dong , f. , eisenstein , d. j. , evans , m. l. , fan , x. , finkbeiner , d. p. , friedman , s. d. , frieman , j. a. , fukugita , m. , gillespie , b. , glazebrook , k. , gray , j. , grebel , e. k. , gunn , j. e. , gurbani , v. k. , de haas , e. , hall , p. b. , harris , f. h. , harvanek , m. , hawley , s. l. , hayes , j. , hendry , j. s. , hennessy , g. s. , hindsley , r. b. , hirata , c. m. , hogan , c. j. , hogg , d. w. , holmgren , d. j. , holtzman , j. a. , ichikawa , s .-
i . , ivezi , . , jester , s. , johnston , d. e. , jorgensen , a. m. , juri , m. , kent , s. m. , kleinman , s. j. , knapp , g. r. , kniazev , a. y. , kron , r. g. , krzesinski , j. , kuropatkin , n. , lamb , d. q. , lampeitl , h. , lee , b. c. , leger , r. f. , lin , h. , long , d. c. , loveday , j. , lupton , r. h. , margon , b. , martnez - delgado , d. , mandelbaum , r. , matsubara , t. , mcgehee , p. m. , mckay , t. a. , meiksin , a. , munn , j. a. , nakajima , r. , nash , t. , neilsen , jr .
, e. h. , newberg , h. j. , newman , p. r. , nichol , r. c. , nicinski , t. , nieto - santisteban , m. , nitta , a. , omullane , w. , okamura , s. , owen , r. , padmanabhan , n. , pauls , g. , peoples , j. j. , pier , j. r. , pope , a. c. , pourbaix , d. , quinn , t. r. , richards , g. t. , richmond , m. w. , rockosi , c. m. , schlegel , d. j. , schneider , d. p. , schroeder , j. , scranton , r. , seljak , u. , sheldon , e. , shimasaku , k. , smith , j. a. , smoli , v. , snedden , s. a. , stoughton , c. , strauss , m. a. , subbarao , m. , szalay , a. s. , szapudi , i. , szkody , p. , tegmark , m. , thakar , a. r. , tucker , d. l. , uomoto , a. , vanden berk , d. e. , vandenberg , j. , vogeley , m. s. , voges , w. , vogt , n. p. , walkowicz , l. m. , weinberg , d. h. , west , a. a. , white , s. d. m. , xu , y. , yanny , b. , yocum , d. r. , york , d. g. , zehavi , i. , zibetti , s. , & zucker , d. b. 2006 , , 162 , 38 , r. j. , kochanek , c. s. , brodwin , m. , brown , m. j. i. , caldwell , n. , cool , r. j. , eisenhardt , p. , eisenstein , d. , gonzalez , a. h. , jannuzi , b. t. , jones , c. , mckenzie , e. , murray , s. s. , & stern , d. 2008 , , 676 , 286 , g. , heckman , t. m. , white , s. d. m. , charlot , s. , tremonti , c. , brinchmann , j. , bruzual , g. , peng , e. w. , seibert , m. , bernardi , m. , blanton , m. , brinkmann , j. , castander , f. , csbai , i. , fukugita , m. , ivezic , z. , munn , j. a. , nichol , r. c. , padmanabhan , n. , thakar , a. r. , weinberg , d. h. , & york , d. 2003 , , 341 , 33 | we have identified a sample of 41 low - mass high oxygen abundance outliers from the mass metallicity relation of star - forming galaxies measured by @xcite .
these galaxies , which have @xmath0 over a range of @xmath1 and @xmath2 , are surprisingly non - pathological .
they have typical specific star formation rates , are fairly isolated and , with few exceptions , have no obvious companions .
morphologically , they are similar to dwarf spheroidal or dwarf elliptical galaxies .
we predict that their observed high oxygen abundances are due to relatively low gas fractions , concluding that these are transitional dwarf galaxies nearing the end of their star formation activity . |
over the past decade , a new solar problem has emerged as solar photospheric abundances have been improved ( @xcite ) and the indicated solar metallicity , which is mainly due to low - z metallic elements , has been significantly revised downward than previously assumed ( @xcite ) .
standard solar models ( ssms ) , which are the fundamental theoretical tools to investigate the solar interior , do not reproduce helioseismic measurements ( @xcite ) , such as the convection zone radius , the surface helium abundance and the sound speed profile ( @xcite ) , when using these revised abundances .
this gave rise to the _ solar composition problem _ , motivating a rapid increase of research efforts in the field .
rosseland opacity is a key quantity describing the coupling between radiation and matter in the solar interior plasma .
metallic elements significantly contribute to the opacity , although they are only present as a few percent of the mixture , since bound - bound and bound - free absorption can become a major source of opacity for the partially ionized metals .
thus , the amount of metals directly modifies the opacity profile of the sun and the solar composition problem is strongly related to the role of opacity in solar models . among a variety of proposed explanations
, it has been shown that in order to reproduce the helioseismic measurements , changes to the opacity are required to compensate the lower solar metallicity .
specifically , it was shown , that a smooth increase of the opacity from the range of @xmath0 in the central regions to the range of @xmath1 at the radiation - convection interface is required ( see @xcite and references therein ) .
it is believed that the opacities of metals in the stellar mixture should be revised upward to compensate for the decreased low - z metallic element abundances .
in addition , recently , the monochromatic opacity of iron , a major contributer to the solar opacity , was measured at very similar conditions to those expected near the solar convection zone boundary ( @xcite ) .
the measured spectrum appears to be larger than calculations of various state of the art and widely used atomic models , by about a factor of two in several spectral regions near the l - shell photoabsorption lines .
no satisfactory explanation to this discrepancy is known .
we have recently developed ( @xcite ) an atomic code , named star ( sta - revised ) , for the calculation of opacities of local thermodynamic equilibrium plasmas , by the sta method .
the sta method ( @xcite ) is used to group the large number of transition lines between an enormous number of configurations into large assemblies , called a super - transition - arrays ( stas ) .
the stas moments are calculated analytically using the partition function algebra , and are split into smaller stas until convergence is achieved . unlike detailed line accounting ( dla ) methods such as the opacity - project ( op ) ( @xcite ) and opal ( @xcite ) ,
the sta method is able to handle situations where the number of relevant transitions is immense . on the other hand , unlike average - atom models ( @xcite ) , the sta method reveals the unresolved - transition - array ( uta ) spectrum , otherwise obtained by a very costly , sometimes intractable , detailed configuration accounting ( dca ) calculation ( see ) . in this work we use star to calculate and
analyze solar opacities .
the atomic calculations are based on a fully relativistic quantum mechanical theory via the dirac equation .
we analyze in detail the contribution of different chemical elements in the solar mixture to the total rosseland mean opacity and examine the role of different atomic processes ( bound - bound , bound - free , free - free and scattering ) .
the analysis is carried throughout the radiative zone ( that is , from the solar core to the radiation - convection boundary ) .
the development of advanced stellar atmosphere models resulted in a downward revision of the solar photospheric metal abundances . in this work
we use the recent @xcite ( agss09 ) set of chemical abundances for volatile elements ( c , n , o , and ne ) together with their recommended meteoritic abundances for refractory elements ( mg , si , s , and fe ) , where si is the anchor of both scales .
this compilation was used as an input for evolutionary solar model calculations presented in @xcite ( and references therein ) , which we refer to as `` the ssm '' . the resulting solar profiles . ] of various quantities
are used throughout this work .
the ssm temperature and density profiles are shown in .
the solar mixture of the ssm consists of the following 24 elements : h , he , c , n , o , ne , na , mg , al , si , p , s , cl , ar , k , ca , sc , ti , v , cr , mn , fe , co and ni . the elemental composition profiles are shown in .
these are determined from the initial abundances and from the ssm due to processes such as nuclear burning , diffusion and gravitational settling .
the opacity used in the ssm was calculated by the opacity - project ( op ) ( @xcite ) .
the effect of intrinsic opacity changes ( that is , with a fixed composition ) on observable quantities , i.e. the convection - zone radius , surface helium abundance , sound speed profile etc , can be studied by the linear solar model ( lsm ) of @xcite .
this justifies the calculation of different opacity models using the thermodynamic conditions obtained by the ssm .
in this section we briefly describe various parts of the computational model implemented by the star atomic code .
a more comprehensive description of the star model will be reported elsewhere .
the initial step in the calculation of the opacity of a mixture is the calculation of the effective densities of the different components ( @xcite ) .
a mixture of temperature @xmath2 , density @xmath3 and mass fractions @xmath4 is described as a collection of ion - spheres with different volumes , or equivalently , individual _ effective mass densities _ @xmath5 .
the individual densities are found by the requirement that the chemical potentials @xmath6 of all species are equal : @xmath7 and by the condition that the volume of the mixture is equal to the sum of the volumes occupied by the individual components : @xmath8 the equality of chemical potentials insures the equality of _ electronic _ pressures only , rather than the total pressure ( including the ionic contribution , see @xcite ) .
however , the usual procedure adopted in the present work , is to equalize the chemical potentials @xcite , which , if the free electrons are treated semiclassically , also insures the equality of the ( free ) electronic densities at the ion sphere radius .
the chemical potential can be calculated using a finite temperature thomas - fermi model ( @xcite ) for which bound and free electrons are treated semiclasically , or alternatively , with more sophisticated ion - sphere models such as the relativistic hartree - fock - slater average atom model for which bound - electrons are treated quantum - mechanically via the dirac equation ( @xcite ) , or the liberman model for which bound and free electrons are treated quantum - mechanically ( @xcite ) . in this work
we used the relativistic hartree - fock - slater average atom model for the calculation of effective mass densities for the solar mixture .
the solar mixture degeneracy parameter @xmath9 ( where @xmath10 is the boltzmann s constant ) is shown in .
it is evident that the degeneracy increases towards the center due to the high increase of matter density , and despite the ( lower ) increase in temperature .
effective mass densities of several elements are given in . .
the monochromatic opacity is evaluated as a sum of four different contributions involving photon scattering ( sc ) and the free - free ( ff ) , bound - free ( bf ) and bound - bound ( bb ) photoabsorption processes . for each component of the mixture
, the photon - absorption ( including stimulated emission ) cross - section is written as : @xmath11 ( 1-e^{-e / k_{b}t } ) .
\end{gathered}\end{aligned}\ ] ] the calculation of the cross - section of each element is carried at the effective density as explained in last paragraph .
the monochromatic opacity is given by the absorption coefficient per unit mass : @xmath12 where @xmath13 is the atomic mass of the @xmath14 component and @xmath15 is avogadro s constant .
the total monochromatic opacity of the mixture is given by the sum of the individual opacities , weighted by the mass fraction : @xmath16 total monochromatic opacities , together with the contribution of individual atomic processes are shown in , for oxygen , magnesium and iron at the conditions of the radiation - convection boundary . the rosseland mean opacity is given by : @xmath17 with the rosseland weight function : @xmath18 where @xmath19 .
we adopt the following simple way to measure fractional contributions ( i.e. the contribution of different elements or different atomic processes ) , to the total rosseland mean .
suppose an opacity spectrum is composed of @xmath20 individual spectral contributions : @xmath21 we define the fractional rosseland contributions by : @xmath22 where @xmath23 is the rosseland mean of the cumulative spectra @xmath24 , while for the first contributer @xmath25 . in this work
we calculate the photon scattering with several corrections as summarized in the book by @xcite .
the photon scattering cross section is written as : @xmath26 where @xmath27 is the average number of free electrons and the thomson cross - section is @xmath28 with @xmath29 the fine structure constant and @xmath30 the bohr radius .
@xmath31 corrects the klein - nishina cross - section due to the finite temperature and contains corrections for inelastic scattering .
it was calculated by @xcite using a relativistic maxwell - boltzmann distribution for _ nondegenerate _ electrons . for nonrelativistic temperatures
it is given by the expansion : @xmath32 where @xmath33 .
the factor @xmath34 that was calculated by @xcite , contains corrections due to the pauli blocking as a result of partial degeneracy and includes a relativistic electron dispersion relation . for @xmath35 ( which holds in the solar interior ) it is given by : @xmath36 \\ & & -\dfrac{345i_{3/2}(\eta)}{128i_{1/2}(\eta ) } \left[1-r^{nr}(\eta)\left ( \dfrac{30i_{3/2}(\eta)}{23i_{1/2}(\eta)}-\dfrac{7i_{5/2}(\eta)}{23i_{3/2}(\eta)}\right ) \right ] \notag t'^{2},\end{aligned}\ ] ] where @xmath37 is the analogous non - relativistic correction due to @xcite : @xmath38 and the fermi - dirac integral is defined by : @xmath39 finally , @xmath40 is a correction due to plasma collective effects .
we use the formula of @xcite that is based on electron correlations at high densities and includes an exchange - correlation correction : @xmath41 the ion and electron debye lengths are given by : @xmath42 where @xmath43 is the vacuum dielectric constant and @xmath44 is the unit electron charge , the total debye length is : @xmath45 and @xmath46 , \end{gathered}\end{aligned}\ ] ] with : @xmath47 we note that for long photon wavelengths where @xmath48 , a direct numerical calculation of eq .
can be unstable . in practice , for @xmath49
it is preferable to use an expansion at @xmath50 : @xmath51 in this work , the free - free photoabsorption is calculated via a screened - hydrogenic approximation with a multiplicative degeneracy correction .
the free - free absorption cross section is given by : @xmath52 where the kramers cross - section ( @xcite ) is : @xmath53 with @xmath54 the free electron number density .
to allow a reasonably accurate and fast calculation of the ( thermal ) averaged free - free gaunt factor @xmath55 , we use a screened - hydrogenic approximation and a maxwellian distribution via the tables of @xcite .
degeneracy is incorporated using the correction factor ( @xcite ) : @xmath56 we note that since @xmath57 depends only on the temperature and the chemical potential , the degeneracy correction is the same for all components in the mixture .
the effect of degeneracy at different solar radii is presented in .
it is evident that the correction is vital at the core vicinity , as the plasma is more degenerate ( see also ) . in principle
, the bound - bound photoabsorption spectra results from all radiative transitions between all levels from all pairs of electronic configurations .
unfortunately , for a hot dense plasma , the number of lines between each pair of configurations may be enormous ( @xcite ) and statistical methods must be used . the unresolved - transition - array ( uta ) method ( @xcite )
treats all levels between each pair of configurations statistically , using analytic expressions for the energy moments of the transition array . in many cases
, the number of configurations may also be extremely large and a detailed - configuration - accounting ( dca ) calculation is intractable as well .
let us estimate the number of populated configurations . a ( relativistic ) configuration is defined by a set of occupation numbers @xmath58 on relativistic @xmath59 orbitals , which are full solutions of the dirac equation .
neglecting electron correlations , the probability distribution of @xmath58 is binomial ( @xcite ) : @xmath60 where @xmath61 is the fermi - dirac distribution , @xmath62 is the orbital degeneracy and @xmath63 is the orbital energy .
the variance of the population in each shell is given by : @xmath64 so that the occupation numbers of shells whose energies are near the fermi - dirac step @xmath65 fluctuate , while the other shells are either filled or empty .
the number of populated configurations @xmath66 can be estimated by the `` width '' of the multivariate distribution : @xmath67 a calculation of @xmath68 for various elements in the solar mixture is shown in .
the sharp jumps are a direct result of the sudden decrease in the number of bound orbitals as they dissolve into the continuum .
this is the process of pressure ionization which is more dominant towards the solar core as the density increases .
it is evident that the number of configurations grows exponentially with the atomic number .
near the convection - radiation boundary , @xmath69 for atomic numbers @xmath70 and a detailed configuration accounting ( dca ) calculation is very costly , while a full detailed - line - accounting ( dla ) calculation is probably impossible .
the bound - bound photoabsorption cross - section is calculated by the sta method ( @xcite ) .
we follow the notation of @xcite for which relativistic @xmath59 orbitals are grouped into
_ supershells _
configurations are collected into large groups called _ superconfigurations _ ( sc ) , defined by a set of occupation numbers @xmath72 over the supershells and denoted by @xmath73 . for each orbital jump @xmath74 and
sc , all transitions between levels in @xmath75 to levels in @xmath76 ( which is obtained from all configurations in @xmath75 via the electron jump ) , are grouped into a single super - transition - array ( sta ) whose first three moments , the intensity , average energy and variance , are respectively : [ eq : moments_sta_def ] @xmath77 @xmath78 @xmath79 where @xmath80 is the energy of level @xmath14 and the line @xmath81 intensity is @xmath82 with @xmath83 the population of level @xmath14 and @xmath84 the line cross - section
. the moments can be expressed as sc averages of occupation number polynomials that can be written in terms of generalized partition functions and evaluated numerically using recursion relations ( @xcite ) , known as the _ partition function algebra_. the ionic distribution @xmath85 where @xmath86 is the number of bound electrons and @xmath87 is the population of the configuration @xmath88 , is also calculated via the partition function algebra .
ion distributions for iron and oxygen are given in and the average ionization for various chemical elements as a function of the solar radius are given in .
it is evident that the ionization increases towards the solar core as a result of the higher temperature and pressure ionization .
the bound - bound spectra contains the contribution of all line profiles from each orbital jump from all scs . in this work
we use voigt profile , and the total bound - bound spectra is written as : @xmath89 the voigt profile @xmath90 is defined by the convolution of the gaussian and lorentzian profiles : @xmath91 the total gaussian width @xmath92 includes the doppler width @xmath93 and the statistical width @xmath94 resulting from the fluctuations in the occupation numbers of the various configurations in @xmath75 and from the uta widths ( @xcite ) .
@xmath95 is the total lorentz width due to the natural and stark broadening effects .
we use the well known semi - empirical formulas for electron impact broadening in the one - perturber approximation by @xcite . in the sta calculation , supershells are split recursively , giving rise to a larger number of scs and a larger number of stas in .
this procedure is repeated until the rosseland mean opacity ( or alternatively , any other criteria ) has converged . for mid or high - z hot dense plasmas ,
the convergence is very fast since the spectral lines are broadened and significantly overlap @xcite .
we note that the supershells are split according to the fluctuations in their occupation numbers @xmath96 , which allows a very robust and fast convergence . the bound - bound sta formalism is straightforwardly generalized for bound - free transitions ( @xcite ) which is presented briefly in this subsection .
the generalization is obtained by treating the continuum orbitals via the fermi - dirac statistics while not including them in the supershell structure .
the total bound - free cross section contains contributions due to photoionization from all bound orbitals @xmath29 : @xmath97 where @xmath98 is given by the convolution : @xmath99 over the free orbital energy @xmath100 .
@xmath101 is given by the sum @xmath102 of all individual sta profiles : @xmath103 -\epsilon , \sigma_{\xi}^{\alpha},\gamma _
{ \alpha } \right),\end{aligned}\ ] ] centered at the energy @xmath104 and include a gaussian width @xmath105 where @xmath106 , @xmath107 and @xmath108 are respectively the doppler , sta and lorentzian widths .
the profile intensity is @xmath109 , where @xmath110 is the total sc population and @xmath111 is the sc average of the number of electrons in orbital @xmath29 .
these moments are easily calculated via the partition function algebra ( @xcite ) analogously to the bound - bound case .
@xmath112 describes the total coupling of the orbital @xmath29 to the continuum via photon absorption : @xmath113 where the summation is over the total and spatial angular momentum quantum numbers of the free orbital @xmath114 and the coefficient @xmath115 is the average number of available `` holes '' in the free orbital .
@xmath116 is the total cross - section for the transition @xmath117 and is calculated via the relativistic expressions of @xcite .
the sum in is restricted by the selection rules of the transition .
the opacity at the solar core calculated with a variety of atomic models was compared by @xcite .
solar opacities by the widely used opal ( @xcite ) and op ( @xcite ) codes were compared in detail by @xcite .
additional calculations and comparisons with the ledcop and atomic codes can be found in @xcite .
average atom solar opacity calculations were presented by @xcite and @xcite .
a recent analysis of the solar opacity was carried by @xcite and @xcite , using the opas detailed configuration accounting model . in this section
we analyze the solar opacity throughout the radiative zone , calculated by star .
the resulting rosseland opacity and mean free path @xmath118 profiles , are shown in .
we note that even though the solar mixture is more opaque at larger radii , the mean free path is larger as well , due to the ( approximate ) exponential decrease in density ( see and ) . the total monochromatic opacity and the monochromatic contributions of various chemical elements , at three different solar radii are shown in , together with the rosseland weight function .
it is evident from the figure that , as is well known , despite their low abundance , metals contribute significantly to the opacity as some of them are not fully ionized and strong bound - bound transition lines and bound - free edges appear nearby the rosseland peak , whose location varies as @xmath119 . in the contribution of various elements to
the rosseland mean opacity along the radiation zone is presented .
these results are in agreement with figure 10 of @xcite and figure 2 of @xcite .
we note that the maxima and minima are caused by the metallic contributions as a direct result of the variations ( mainly due to the varying temperature ) in the positions of lines and edges relative to the position of the rosseland peak .
for example , at the solar core or near the convection - radiation boundary , the iron k and l shell features are , respectively , nearby the rosseland peak , while at @xmath120 they are both well outside the rosseland distribution ( as seen in ) .
this results in a peak near the core , a minimum near @xmath120 and a peak near the radiation - convection interface , in the iron opacity contribution profile , as shown in .
a complete plot of orbital energies relative to @xmath121 for the @xmath122 , @xmath123 and @xmath124 orbitals and for various elements across the radiation zone is shown in .
the relevance of each element due to each of its orbitals is inferred from the proximity to the rosseland peak .
it is also evident that orbital energies become less negative towards the core since the screening of the nucleus is increasing due to the increase in density . as a result
, some orbitals can be pressure ionized into the continuum at a finite radii .
indeed , as seen in , the @xmath124 orbital of hydrogen and helium , the @xmath123 orbital of c , n , o , ne , mg , al and si and the @xmath122 orbital of all elements in the solar mixture , pressure ionize at a finite radii .
the contribution of the bound - bound and bound - free processes to the rosseland opacity depends on the positions of photoabsorption lines and edges relative to the position of the rosseland peak and on the population of the various bound states .
relative contributions of the four different atomic processes : scattering , free - free , bound - free and bound - bound , to the elemental rosseland opacity of various elements throughout the radiation zone , are presented in .
the results are in agreement with figure 2 of @xcite .
the behavior of the bound - bound and bound - free contributions in can be qualitatively understood from and . for hydrogen and helium ,
all orbitals are pressure ionized up to some finite radii ( @xmath125 and @xmath126 , respectively ) and the bound - free contribution is zero up these threshold radii . in addition , orbitals with principle quantum number @xmath127 are pressure ionized throughout the radiation zone so that there is no bound - bound contribution for hydrogen and helium . on the other hand , for all other elements the k - shell exists throughout the radiation zone and
the bound - free contribution is always non - zero .
however , several elements have a finite threshold radius for the existence of the l - shell orbitals and therefore a finite threshold radius for the bound - bound contribution ( see fig .
[ fig : fracs_dk_elements]c to fig.[fig : fracs_dk_elements]k ) .
this threshold radius is lower for heavier elements , for which the @xmath128-shell orbitals are closer to the heavier nucleus and a larger density is required for pressure ionization . for the heaviest elements ( fig .
[ fig : fracs_dk_elements]l to fig.[fig : fracs_dk_elements]o ) , the l - shell exists throughout the radiation zone and therefore , they always have a non - zero bound - bound contribution .
as we have mentioned , as the temperature decreases outward , bound - bound lines move with respect to the rosseland peak ( ) .
k - shell lines of certain light elements , only approach the rosseland peak from below , explaining the monotonic behavior of their bound - bound contributions ( fig .
[ fig : fracs_dk_elements]c to fig.[fig : fracs_dk_elements]f ) , while heavier elements k - shell lines cross the rosseland peak and have a local maximum in their bound - bound contribution ( fig .
[ fig : fracs_dk_elements]g to fig.[fig : fracs_dk_elements]j ) . for heavier elements , after their k - shell lines pass the rosseland peak , at large enough radii , their l - shell lines also begin to approach it , giving rise to local maximum followed by a local minimum in the bound - bound contribution ( fig .
[ fig : fracs_dk_elements]k to fig.[fig : fracs_dk_elements]m ) . finally , we note that the iron and nickel k - shell lines at @xmath129 are already past the rosseland peak , while their l - shell lines cross it near the radiation - convection boundary , giving rise to a local minimum followed by a local maximum ( fig .
[ fig : fracs_dk_elements]n and fig.[fig : fracs_dk_elements]o ) .
the total contribution of the different atomic processes ( due to all elements combined ) to the ssm mixture rosseland opacity is shown in .
we compare star and opacity - project ( op ) monochromatic opacities and ion charge state distributions for o , ne , mg and fe , which have a major opacity contribution near the radiation - convection interface .
comparisons are made at @xmath130 and @xmath131 available in the topbase database ( @xcite ) , which is used to generate op data .
the same comparisons were made by @xcite for fe and mg with opas and by @xcite for fe and o using the ledcop and atomic opacity codes .
the ion distributions are given in figures [ fig : pq_op_o]-[fig : pq_op_fe ] .
the opas charge distributions given in @xcite for mg and fe are also shown in figures [ fig : pq_op_mg]-[fig : pq_op_fe ] . for o and ne ,
mean ionizations @xmath132 of op and star are in excellent agreement , though the charge distributions slightly differ .
we note that for fe , larger differences appear and for the lower charge states ( @xmath133 ) , star populations are much larger . in these lower ionic states , ground and excited configurations must have a non - empty m - shell .
this tendency is similar to that shown by opas , though the star populations are somewhat larger than opas at the lower charge states .
a comparison between op and star o , ne , mg and fe monochromatic opacities are given in figures [ fig : op_specs_o ] , [ fig : op_specs_ne ] , [ fig : op_specs_mg ] and [ fig : op_specs_fe ] , respectively .
the rosseland means are also given in the figures .
it is evident that op k - shell lines that are shown for o , ne and mg , are significantly broader than the star lines .
the broad op k - shell wings which lie nearby the rosseland peak , contribute significantly to the rosseland mean , as seen in figures [ fig : op_specs_o ] , [ fig : op_specs_ne ] , [ fig : op_specs_mg ] .
such differences were also observed by @xcite for mg and by @xcite for o. both star and op line width calculations include electron impact broadening in the one - perturber approximation .
however , op uses collisional rates obtained from comprehensive quantum calculations that include unitarization ( @xcite ) , while star uses the widely - used semi - empirical formulas by @xcite .
a comparison of the iron monochromatic opacity is shown in .
it is evident that the star opacity is larger than op near the m - shell bound - bound and bound - free regions , which contribute significantly to the rosseland mean .
similar differences were also found by @xcite .
this larger m - shell contribution is explained by the larger star population of the lower ionic states ( ) , for which ground and excited configurations have a non empty m - shell .
the sta method takes into account the huge number of excited configurations in these lower ion states by the robust superconfiguration approach .
a comparison of star and opas with op rosseland mean opacities at @xmath134 and @xmath135 for the @xmath136 elements available in the topbase database are given in ( the opas data was taken from @xcite ) . for ne , na , mg , al , and si
, the op opacity is significantly higher than star due to the larger widths of the k - shell lines . on the other hand , for cr , mn , fe and ni , star opacity is significantly larger than op due to higher population of non - empty m - shell configurations .
it is also evident that the star differences are in a relatively good agreement with those of opas , except for carbon for which the star agrees better with op .
a comparison of the solar opacity profile throughout the radiation zone , between calculations via star and opal ( given by @xcite ) , to the opacity profile calculated by op ( @xcite ) , is presented in .
it can be seen that the star and op calculations agree within @xmath137 throughout the radiation zone .
the solar mixture opacity is quite insensitive to the individual element opacities , so that , as was shown by @xcite , despite the large differences between several elemental opacities ( ) , the total mixture star opacity is in good agreement with op .
this is possible since for ne , na , mg , al and si , star opacities are significantly lower than op , while for cr , mn , fe and ni , star opacities are significantly higher than op .
metals heavier than nickel have a very low abundance in the solar mixture , and therefore are usually neglected in ssms .
however , at solar conditions , heavy metals may have a large number of bound electrons and therefore their spectra usually contain a huge number of strongly absorbing lines , which may in principle , affect the solar opacity .
due to the huge number of lines , a tractable calculation of hot high - z elements spectra is only possible using the sta or average - atom methods .
solar opacities including heavier elements from the @xcite composition were calculated and analyzed by @xcite using the sta program of @xcite .
it was shown that even though the heavy elements are very strong photon absorbers , their very low abundance leads to a negligible effect on the rosseland mean opacity . here
we consider the rosseland mean opacity of the solar mixture including _ all _ elements in the recent agss09 ( @xcite ) photospheric composition .
the calculation is done at @xmath138 and @xmath139 that exist nearby the radiation - convection interface . in the mass and number fractions of all elements in the agss09 mixture
are given , together with the elemental contribution to the mixture rosseland mean opacity .
the table is also visualized in .
it is evident that the heavy metals opacity fractions are larger by more than four ( two ) orders of magnitude than their number ( mass ) fractions , indicating that heavy metals are indeed good absorbers .
however , as was also shown by @xcite , it is seen that due to their extremely small abundance , the effect of the heavy metals on the total rosseland mean opacity is completely negligible .
the contribution of the different atomic processes for the various elements are given in .
we note that the local maxima in the bound - bound contribution represent the @xmath140 and @xmath141 shell lines crossing of the rosseland peak ( here , the crossing occurs due to the dependence of line energies on the atomic number ) .
the estimated number of populated configurations ( ) of the different elements is given in .
it is evident that according to the estimate , for mid and high - z metals , a complete dca ( and of course , dla ) calculation is absolutely intractable , and sta or average - atom ( @xcite ) methods must be used .
@lllll@ @xmath142 & element & .specification of the agss09 photospheric admixture mass and number fractions .
elemental contribution to the rosseland opacity , calculated by star at @xmath143 that are found nearby the radiation - convection interface , are also listed .
@xmath144 represents @xmath145 [ cols= " < " , ] + @xmath146 & h & @xmath147 & @xmath148 & @xmath149 + @xmath150 & he & @xmath151 & @xmath152 & @xmath153 + @xmath154 & li & @xmath155 & @xmath156 & @xmath157 + @xmath158 & be & @xmath159 & @xmath160 & @xmath161 + @xmath162 & b & @xmath163 & @xmath164 & @xmath165 + @xmath166 & c & @xmath167 & @xmath168 & @xmath169 + @xmath170 & n & @xmath171 & @xmath172 & @xmath173 + @xmath174 & o & @xmath175 & @xmath176 & @xmath177 + @xmath178 & f & @xmath179 & @xmath180 & @xmath181 + @xmath182 & ne & @xmath183 & @xmath184 & @xmath185 + @xmath186 & na & @xmath187 & @xmath188 & @xmath189 + @xmath190 & mg & @xmath191 & @xmath192 & @xmath193 + @xmath194 & al & @xmath195 & @xmath196 & @xmath197 + @xmath198 & si & @xmath199 & @xmath200 & @xmath201 + @xmath202 & p & @xmath203 & @xmath204 & @xmath205 + @xmath206 & s & @xmath207 & @xmath208 & @xmath209 + @xmath210 & cl & @xmath211 & @xmath212 & @xmath213 + @xmath214 & ar & @xmath215 & @xmath216 & @xmath217 + @xmath218 & k & @xmath219 & @xmath220 & @xmath221 + @xmath222 & ca & @xmath223 & @xmath224 & @xmath225 + @xmath226 & sc & @xmath227 & @xmath228 & @xmath229 + @xmath230 & ti & @xmath231 & @xmath232 & @xmath233 + @xmath234 & v & @xmath235 & @xmath236 & @xmath237 + @xmath238 & cr & @xmath239 & @xmath240 & @xmath241 + @xmath242 & mn & @xmath243 & @xmath244 & @xmath245 + @xmath246 & fe & @xmath247 & @xmath248 & @xmath249 + @xmath250 & co & @xmath251 & @xmath252 & @xmath253 + @xmath254 & ni & @xmath255 & @xmath256 & @xmath257 + @xmath258 & cu & @xmath259 & @xmath260 & @xmath261 + @xmath262 & zn & @xmath263 & @xmath180 & @xmath264 + @xmath265 & ga & @xmath266 & @xmath267 & @xmath268 + @xmath269 & ge & @xmath270 & @xmath271 & @xmath272 + @xmath273 & kr & @xmath274 & @xmath275 & @xmath276 + @xmath277 & rb & @xmath278 & @xmath279 & @xmath280 + @xmath281 & sr & @xmath282 & @xmath283 & @xmath284 + @xmath285 & y & @xmath286 & @xmath287 & @xmath288 + @xmath289 & zr & @xmath290 & @xmath291 & @xmath292 + @xmath293 & nb & @xmath294 & @xmath295 & @xmath296 + @xmath297 & mo & @xmath298 & @xmath299 & @xmath300 + @xmath301 & ru & @xmath302 & @xmath303 & @xmath304 + @xmath305 & rh & @xmath306 & @xmath307 & @xmath308 + @xmath309 & pd & @xmath310 & @xmath311 & @xmath312 + @xmath313 & ag & @xmath314 & @xmath315 & @xmath316 + @xmath317 & in & @xmath318 & @xmath319 & @xmath320 + @xmath321 & sn & @xmath322 & @xmath323 & @xmath324 + @xmath325 & xe & @xmath326 & @xmath327 & @xmath328 + @xmath329 & ba & @xmath330 & @xmath331 & @xmath332 + @xmath333 & la & @xmath334 & @xmath335 & @xmath336 + @xmath337 & ce & @xmath338 & @xmath339 & @xmath340 + @xmath341 & pr & @xmath342 & @xmath343 & @xmath316 + @xmath344 & nd & @xmath345 & @xmath346 & @xmath347 + @xmath348 & sm & @xmath267 & @xmath349 & @xmath350 + @xmath351 & eu & @xmath352 & @xmath353 & @xmath354 + @xmath355 & gd & @xmath356 & @xmath357 & @xmath358 + @xmath359 & tb & @xmath360 & @xmath361 & @xmath362 + @xmath363 & dy & @xmath364 & @xmath335 & @xmath365 + @xmath366 & ho & @xmath367 & @xmath368 & @xmath369 + @xmath370 & er & @xmath371 & @xmath372 & @xmath373 + @xmath374 & tm & @xmath375 & @xmath376 & @xmath377 + @xmath378 & yb & @xmath379 & @xmath380 & @xmath381 + @xmath382 & lu & @xmath383 & @xmath376 & @xmath384 + @xmath385 & hf & @xmath386 & @xmath387 & @xmath388 + @xmath389 & w & @xmath390 & @xmath387 & @xmath391 + @xmath392 & os & @xmath393 & @xmath394 & @xmath395 + @xmath396 & ir & @xmath397 & @xmath160 & @xmath398 + @xmath399 & au & @xmath400 & @xmath372 & @xmath401 + @xmath402 & tl & @xmath403 & @xmath404 & @xmath204 + @xmath405 & pb & @xmath406 & @xmath303 & @xmath407 + @xmath408 & th & @xmath409 & @xmath410 & @xmath411 +
a new opacity code , star , implementing the sta method of @xcite , was presented .
it was used to calculate and analyze opacities throughout the radiation zone for a solar model implementing the recent agss09 composition .
the relative contribution of various elements and atomic processes to the rosseland mean opacity were compared in detail .
star and op monochromatic opacities and charge state distributions were compared for several elements near the radiation - convection interface . for the heavier elements , such as iron ,
star rosseland opacity is significantly larger than op , which is explained by the larger population of lower ionization states . for the lighter elements , such as magnesium ,
op rosseland opacity is significantly larger than star , due to a much larger line broadening of the k - shell lines shown by op .
very similar results were reported by @xcite using the opas code .
star rosseland opacities for the solar mixture were calculated throughout the radiation zone and a good agreement with op and opal was achieved . finally , an explicit star calculation of the full agss09 photospheric mixture , including all heavy metals was performed .
it was shown that due to their extremely low abundance , and despite being very good photon absorbers , the heavy elements do not affect the rosseland opacity , in agreement with @xcite .
+ + the authors thank aldo serenelli and francesco villante for providing various solar profiles and for useful suggestions and comments . | a new opacity model based on the super - transition - array ( sta ) method for the calculation of monochromatic opacities of local thermodynamic equilibrium plasmas , was developed .
the atomic code , named star ( sta - revised ) , is described and used to calculate spectral opacities for a solar model implementing the recent agss09 composition .
calculations are carried throughout the solar radiative zone .
the relative contributions of different chemical elements and atomic processes to the total rosseland mean opacity are analyzed in detail .
monochromatic opacities and charge state distributions were compared with the widely used opacity - project ( op ) code , for several elements near the radiation - convection interface .
star rosseland opacities for the solar mixture show a very good agreement with op and the opal opacity code , throughout the radiation zone .
finally , an explicit sta calculation of the full agss09 photospheric mixture , including all heavy metals was performed .
it was shown that due to their extremely low abundance , and despite being very good photon absorbers , the heavy elements do not affect the rosseland opacity . |
the next generation cherenkov telescope array ( cta ) will be constituted of three types of imaging atmospheric cherenkov telescopes : small- , medium- and large - sized telescopes .
the array will be dedicated to the observation of the high energy gamma - ray sky with unprecedented sensitivity @xcite over a broad range of energies ( 0.01tev@xmath1100tev ) .
this instrument will enable astronomers and astro - particle physicists to refine models of gamma - ray sources and underlying non - thermal mechanism at work , to study the origin and the composition of the cosmic rays up to the knee region , question the nature of the dark matter , etc .
these kind of telescopes detect cherenkov light emitted along developing showers in the atmosphere nearly uniformly illuminating the ground over an area of about 50000m@xmath0 .
shower maxima of a gamma - ray induced electromagnetic shower occur at altitudes comprised between @xmath27 km and @xmath212 km , for a primary energy ranging from @xmath2100gev to @xmath2100tev .
the photon density on the ground depends on the energy of the gamma ray , its incidence direction and the distance from the resulting shower axis .
the displacement of the image centroid in the telescope camera is a function of the angular distance of the shower core , defined by shower impact distance and altitude of the shower and is typically about 1.5 for an energy of 1tev and an impact parameter of 150 m , and 3 to 3.5 for energies around 100tev with an impact parameter of 300 m .
as very high energy showers penetrate deeply in the atmosphere and generate a large amount of light , they can be observed at a relatively large distance from the main light pool already with relatively small reflectors translating into the necessity of a deployment of telescopes with large field - of - view for the exploration of very high energy showers . in general , the required telescope field - of view in order to record contained shower images ranges between 3 and 10 depending on the energy range of interest .
a promising design for wide field observations is the davies - cotton telescope @xcite .
its advantage is a point - spread function enabling a larger field - of - view than a parabolic design .
the schwarzschild - coud design ( e.g. @xcite ) is not considered further , as it presents technical and cost challenges compared to a conservative cta proposal using a prime optics design ( see also the comments in section [ sct : winstoncone ] ) .
it is demonstrated that once the field - of - view of the camera pixels and of the whole instrument together with the photo - sensor technology is fixed by means of physics arguments , only one parameter is left free .
this parameter can be fixed as well , relating the worst optical resolution in the field - of - view , i.e.at the edge of the field - of - view , with the size of a pixel . having a reasonable cost model at hand
, even the most cost - efficient single telescope or telescope design operating in an array can be derived . knowing the anticipated angular size of a pixel from physics constraints , a requirement on the point - spread function of the telescope can be determined .
if the point - spread function is identified by the root - mean - square of the light distribution , requiring a pixel diameter four times the root - mean - square ensures that most of the light from a point source at infinity is concentrated in a single pixel at the edge of the field - of - view .
if simulations show that such a containment is not necessary in terms of angular resolution of the shower origin or background suppression , also a smaller value like twice the root - mean - square can be considered , c.f .
if this requirement is not optimized or not met , either the physics outcome is worsened or the camera has more pixels than necessary and will not be cost - efficient .
the cta array layout will consist of large size telescopes ( lst , primary mirror diameter @xmath224 m ) in the center , and successively surrounded with an increasing number of medium - sized ( mst , @xmath212 m ) and small - sized ( sst , @xmath26 m ) telescopes in order to instrument a ground surface area comprised between 4km@xmath0 and 10km@xmath0 .
telescope spacings , sizes and field - of - views reflect the energy range to be explored by a certain type of telescope
. the lst sub - array will be primarily focusing on the observation of the high energy gamma - ray sky with great precision below about 100gev , the inter - telescope spacing will be small , about 60 m and the single telescope field - of - view limited to about 34 .
the sst sub - array will conversely be optimized for multi - tev observations , up to around 100tev allowing for large spacing of up to 300m500 m with relatively modest reflector size .
the camera pixel field - of - view of these different telescope types will span between about 0.080.1 ( lst ) up to 0.20.3 ( sst ) , linked to the different shower intensity with its intrinsic fluctuations and the concurrent necessity of keeping the number of camera pixels reasonably low .
consequently , this study will focus on reflector diameters in the range of a meter to about 30 m and on pixel field - of - views in the range between @xmath20.08 and @xmath20.3 . and is extended to off - axis angles of the incoming light up to more than 5 . in the following ,
first a description is derived for the relations between the existing design parameters of a cherenkov telescopes , e.g.focal length and reflector diameter .
then , by including a semi - analytical treatment for the optical quality of a generalized davies - cotton reflector , this description becomes applicable for the design of real telescopes . at the end , the influence of the variation of the parameters on the optimized design is discussed .
the light collection is one of the most important parameter of a cherenkov telescope and can be improved by , e.g. , an increase of the photo - detection efficiency of the photo - sensors or a rescaling of the system , i.e. , a corresponding rescaling of the reflective area and photo - sensor size . as current technology ( photo - multiplier tubes , hybrid photo - detectors or silicon photo - multipliers ) only offer a limited choice of photo - sensor sizes , the cherenkov telescope design parameter phase space is reduced .
given a particular photo - sensor type , the addition of a light concentrator in front of the photo - sensor is the only way to increase the light collection efficiency . in the following
, it will be shown that the relationship between the characteristic parameters of an optimal telescope design ( pixel field - of - view and linear size , reflector diameter and focal distance ) is fully constrained , once the technological requirement ( photo - sensor size and the light - guide material ) and the physics requirements ( the pixel angular size and the field - of - view ) are frozen .
this result is obtained by imposing restrictions on the value of the point - spread function at the edge of the field - of - view in order to keep adequate image quality and thus analysis potential .
the theorem of liouville states that the maximum concentration theoretically achievable is defined by maintaining the phase space , i.e.the product of solid angle , defined by the incoming light rays directions , light ray momentum squared and the surface area crossed by the light rays .
the theorem of liouville is applicable to the case of a winston cone with entrance area @xmath3 placed in front of a photo - detector of area @xmath4 , corresponding to the exit area of the cone .
@xmath3 is defined provided @xmath4 , the solid angle @xmath5 defined by the light rays entering the cone and the solid angle @xmath6 defined by the light rays leaving the cone .
winston has shown in @xcite that the maximum concentration factor for a rotationally symmetric light concentrator is @xmath7 where @xmath8 and @xmath9 denote the refractive index of the media in front of and inside the optical system with @xmath8@xmath101 in air .
@xmath11 corresponds to the maximum angle at which a light ray enters the system related to the solid angle @xmath5
. if the system is not axisymmetric or the angular acceptance of the photo detector is smaller than @xmath12 ( as assumed in eq . [
eqn : concentration ] ) , @xmath13 has to be adapted accordingly . besides the increase of the light collection area ,
the use of cones enables a partial screening of the night - sky light pollution corresponding to @xmath11 being larger than the angle of light rays coming from the edge of the reflector .
simulations @xcite and recent concentration efficiency measurements @xcite of solid cones ( @[email protected] ) designed for the fact camera @xcite demonstrated that their shape is nearly ideal , that the concentration factor reaches a value close to @xmath13 and the geometric loss is only of the order of a few percent excluding absorption loss . in the case of hollow cones ,
fresnel reflection losses have to be considered at the surface of the photo - sensor .
if the camera is sealed with a protective window , which is usually the case , also losses at the window surface need to be taken into account . by choosing a material for the cones and the protective window with a similar refractive index than the material of the photo - sensor light entrance , these losses can be omitted .
combined with the almost perfect reflectivity of solid cones due to total reflection ( limited only by the surface roughness ) , solid cone usually outperform hollow cone .
the concentration factor achieved with the winston cones is fundamentally similar to the size reduction of the focal plane in a schwarzschild - coud design and linked to the conservation of the space - momentum phase space according to the liouville theorem : the conversion of the spatial into momentum phase space , by means of winston cones or secondary optics . while cones reduce the acceptance of the incoming light rays from a large area at the entrance to a large angular acceptance and small area at the cone exit , the secondary mirror optics leads to a similar spatial compression and angular widening of the light rays at the photo - sensor and thus a reduced plate - scale . while cones are non - imaging devices , the secondary optics is imaging .
hence , in the schwarzschild - coud design , it is possible to attain excellent optical resolution , in terms of cherenkov telescope requirements , with a field - of - view as large as 15 .
both designs enable compression of the photo sensitive area by factors larger than ten w.r.t .
their primary optics design . in the case of a cherenkov telescope ,
the light entering the cone comes from a reflector visible under a maximum angle @xmath11 defined from the focal plane center ] .
the opening angle of the light at the entry of the cone is therefore well defined by the properties of the optical system , i.e.by the diameter of the reflector @xmath14 and the focal length @xmath15 , @xmath16 . @xmath17 combining this with eq .
[ eqn : concentration ] yields @xmath18 for instance , taking the fact values , @xmath19 and @xmath20 , we obtain the theoretically maximal achievable concentration factor @xmath21 , i.e.the linear size of the entrance area can ideally be larger than four times the linear photo - sensor size .
in addition , the optical system defines the zoom factor or plate - scale , i.e.the field - of - view corresponding to a physical area in the focal plane .
the correspondence between the angular size @xmath22 and the linear size @xmath23 on the focal plane is @xmath24 or in the limit of small @xmath22 , @xmath25 cameras in cherenkov telescopes are pixelized due to the use of photo - detectors . to increase the light collection efficiency further , and to maintain symmetry ,
these pixels are usually aligned on a hexagonal grid , i.e.in closed package geometry . in recent years
, magic has exploited the photon arrival time extracted from the measured pulse and demonstrated significant improvements in the sensitivity @xcite .
the technique , taking into account the change of the arrival time between neighboring pixels , performs best , if all neighbors are at an identical distance from the central pixel .
consequently , the ideal shape of a pixel is hexagonal .
the distance @xmath23 on the camera surface is the distance between two parallel sides of a hexagon , its area is @xmath26 combining with eq .
[ eqn : baseformula ] , the plate - scale formula eq .
[ eqn : platescale ] and eq .
[ eqp : areahexagon ] , @xmath27 is obtained , which translates the close relationship between the pixel field - of - view @xmath22 , the focal length @xmath15 and the reflector diameter @xmath14 , once the technological parameters fixed : photo detector size @xmath4 and light concentrator material @xmath9 . defining a constant related to these properties @xmath28 and rewriting eq .
[ eqn : derivedformula ] ( @xmath29 for @xmath301 ) as @xmath31 it is immediately apparent that the focal length @xmath15 of the system is a direct consequence of the pixel field - of - view and the reflector diameter , if the properties of the photon detector and the material of the cones are known . for typical cherenkov telescopes , @xmath32 is between unity and two .
below unity the resolution becomes too coarse and above two , not only the number of pixels and hence the price of a camera becomes too high , but also the camera holding structure becomes mechanically complex and hence disproportionally expensive .
this constraint on @xmath32 applied to eq .
[ eqn : focallength ] yields @xmath33 precisely the choice of @xmath15 and @xmath14 defines the optical quality of a mirror system . at the same time
, the size of a single pixel defines a natural constraint on the optical quality of a system , i.e. , @xmath32 should be chosen such that the point - spread function at the edge of the camera is within a limit well defined by the pixel s field - of - view .
the light collection area is important for a cherenkov telescope and typical reflector sizes range from a few to ten or twenty meters .
however , with the current technology , it is not possible to produce large mirrors with the requested quality at a reasonable cost .
furthermore , optical systems compiled from a single mirror suffer large aberration effects at large off - axis angles , while a wide field - of - view is necessary for the observation of multi - tev showers up to large impact parameters , as well as for extended sources .
therefore , segmented mirrors are in use .
the layout providing the best optical quality for segments of identical focal length is the so - called davies - cotton layout @xcite , where the single spherical mirrors are located on a sphere with radius @xmath15 and focused to a point at @xmath34 . the relevant quantity which influences the on - axis and off - axis optical quality is the focal ratio @xmath32 .
the optical quality improves with larger values .
this scale invariance statement is true only as long as the optical quality of a single mirror can be neglected against the optical quality of the whole system , which is generally the case at the edge of the camera . to be able to constrain the optical point - spread function , a relation between the tessellation , the focal ratio and the resulting point - spread function is needed for a given maximum inclination angle of the light , i.e.at the edge of the field - of - view of the camera . in [ appendix : a ] , a formalism describing the point - spread function of an ideal davies - cotton reflector , i.e.a reflector with infinite tessellation is presented .
the point - spread function is described by the root - mean - square of the light distribution . with the help of ray - tracing simulation , a reasonably good description of a real tessellated davies - cotton reflector
is derived from this analytical approach in [ appendix : b ] . including the correction factor which describes the deviation of the analytical approach from the simulations , a good description of the point - spread function
is obtained .
it is shown that the point - spread function @xmath35 of an ideal davies - cotton can be expanded into a polynomial in @xmath36 .
given the incident angle @xmath37 of the incoming ray and @xmath38 in the range between one and two , the polynomial is hypothesized into the more simpler form @xmath39 with coefficients @xmath40 and @xmath41 and @xmath42 the tessellation number as described in [ appendix : b ] .
this parametrization is found to match the ray - tracing simulation without loss of precision .
the coefficient @xmath40 can directly be deduced as the result of eqs .
[ eq : realdc ] at @xmath43 and @xmath41 is derived by a fit .
an example for the coefficients @xmath40 and @xmath41 for selected @xmath44 is shown in fig .
[ fig : coefficients ] .
as discussed in the introduction , it is required that the point - spread function is small compared to the pixel field - of - view at the edge of the field - of - view , so that the light of a point source is well contained in one pixel . defining the ratio @xmath45 between both , this requirement can be expressed as @xmath46 combined with eq .
[ eqn : pointspreadfunction ] , the focal ratio @xmath38 can now be expressed as @xmath47 including this in eq .
[ eqn : focallength ] yields @xmath48 with a correction factor @xmath49 defined as @xmath50 the absolute focal length @xmath15 can now be calculated using eq .
[ eqn : focallength ] . to deduce the effective reflective area from eq .
[ eqn : result ] , the shadow of the camera on the reflector has to be taken into account .
to calculate the fraction of the reflector shadowed , the ratio of their areas is calculated .
conversion of @xmath44 from an angle to a length yields approximately @xmath51 for small values of @xmath44 ( @xmath5210 ) . expressing the focal length by eq .
[ eqn : focallength2 ] the fraction of the camera shadow on the reflector is derived .
@xmath53 if a real camera housing is significantly larger than the photo sensitive area itself , a correction factor should be included .
now the effective light collection area of the optical system can be deduced as @xmath54 if real setups should be compared , like e.g.davies-cotton and schwarzschild - coud , also other sources of light - losses must be included , such as geometrical efficiency of the cones ( light - loss at the edge of the mirror ) , total mirror reflectivity , cone transmission or reflection losses , or photo detection efficiency . the relation given in eq .
[ eqn : result ] includes several parameters which are subject to change . for simplicity , a standard setup has been defined to which altered setups are compared .
silicon photo - detectors are a recent and very promising technology .
therefore , a silicon photo - detector with a sensitive area of 36mm@xmath0 is chosen as a benchmark device .
such devices are already commercially available with acceptable properties .
although their sensitive area is still rather small compared to photo - multipliers , by increasing their light - collection with solid light concentrators , their light - collection area becomes reasonably large .
such light - concentrators still maintain a reasonable weight and length in term of absorption .
typical plexiglas materials have a refractive indices of the order of @xmath9=1.4 and are used hereafter as a reference .
as the light - collection area of a telescopes scales directly with the photo - sensitive area , the most obvious use of small photo sensors is a small telescope sensitive mostly to high energetic showers . at high energies ,
the collection area of a telescope array is of prime importance due to rapidly decreasing fluxes . due to the bright light - pool of high energy showers , telescopes with relatively small reflectors
can be operated with large spacing of , e.g. , 400 m or more , c.f . @xcite . since such spacings demand a large camera field - of - view , a field - of - view of 9 diameter
is chosen as a reference .
a typical reflector for a cherenkov telescope with davies - cotton layout enables the manufacturing of a primary reflector tessellated into spherical mirror of identical focal lengths . from the scaling with the tessellation number as derived in [ eq : realdc ]
, it can be concluded that a layout with only three mirrors on the diagonal ( @xmath42=3 ) has still a significantly worse optical quality than a reflector with five mirrors on the diagonal .
although the point - spread function at the center of the camera is clearly dominated by the mirror size , the relative influence almost vanishes at higher off - axis angles . since the solution with @xmath42=3 still shows a degradation of more than 10% compared to the solution with @xmath42=5 even at the highest simulated off - axis angles , it is discarded .
on the other hand , a further increase of the tessellation number ( individual mirror size over primary reflector diameter ) does not significantly improve the optical quality .
consequently , choosing @xmath42=5 is a good compromise and already close to the optimum achievable .
comparable results were obtained in @xcite although using a third order approximation overestimating the optical quality .
it must be noted that the simulation does not take the point - spread function of the individual mirrors nor any possible misalignment into account which must be added quadratically to the result .
however , for the solutions discussed here this can be neglected , c.f .
in general , alignment errors can be kept minimal and individual mirrors can be machined with a point - spread function small compared to the point - spread function at the edge of the camera . on average ,
all davies - cotton designs with a reasonable @xmath32 have a root - mean - square of the light distribution in the tangential direction about two times larger than in the sagittal direction .
ideally , the sagittal root - mean - square at the edge of the camera should fit a fourth of the pixel s field - of - view .
this ensures that in the sagittal direction 95% of the light is contained within one pixel diameter and roughly 68% in the tangential direction .
however , since the point - spread function is not gaussian and has long tails in tangential direction exact numbers for the light content might slightly differ . for convenience ,
all following plots show dots for @xmath32=\{1 , 1.25 , 1.5 , 1.75 , 2}. fig .
[ fig : resulta ] shows the reflective area versus the pixel s field - of - view for comparison in the standard case with and without shadowing for different camera field - of - views .
since the effect is comparably small and the mirror diameter is more expressive , in the following all plots show the mirror diameter rather than the reflective surface in the non - obstructed case .
the effects of changing different input parameters w.r.t . to the previously described benchmark configuration
are shown in fig .
[ fig : comparison ] and discussed below .
[ [ changing-the-camera-field-of-view-fig.figcomparison-top-plot ] ] changing the camera field - of - view ( fig .
[ fig : comparison ] , top plot ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + changing the camera s field - of - view basically shifts the valid range along the line , i.e.the range corresponding to @xmath32=[1.0 , 2.0 ] .
that means that it is possible to build telescopes identical in optical quality , pixel s field - of - view and mirror diameter , but different field - of - view resulting simply in a different focal length of the system . in short : changing the field - of - view only changes the focal length .
[ [ changing-the-optical-quality-fig.figcomparison-middle-plots ] ] changing the optical quality ( fig .
[ fig : comparison ] , middle plots ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a change in the requirement on the optical quality @xmath45 directly influences @xmath32 , and therefore also shifts the range of reasonable @xmath32 almost linearly in @xmath22 ( left plot ) .
changing the tessellation ( right plot ) is like changing the requirement on the optical quality .
while the difference in optical quality between a davies - cotton layout with three mirrors on the diagonal and five mirrors is still significant , all other layouts give identical results within a few percent . in short : any tessellation number @xmath555 gives similar results .
changing the requirement on the optical quality only changes the focal length .
[ [ changing-the-photo-sensitive-area-fig.figcomparison-bottom-left-plot ] ] changing the photo sensitive area ( fig .
[ fig : comparison ] , bottom left plot ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + since the constant @xmath56 is directly proportional to the size @xmath4 of the photon detector , the mirror area is directly proportional to the size of the photo sensor .
if the size of the photon sensor is limited , a simple way to increase the field - of - view of a single pixel is to sum the signal of several photon counters to a single signal . to maintain a hexagonal , i.e.most symmetric layout , summing
the signal of three , four or seven photon sensors seems appropriate . in short : assuming an optimized light - concentrator , the photo sensor s physical size defines the scale of the system .
[ [ changing-the-light-concentrator-fig.figcomparison-bottom-right-plot ] ] changing the light concentrator ( fig .
[ fig : comparison ] , bottom right plot ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + another way to increase the reflective area is an increase of the refractive index of the light concentrator entering quadratically .
using solid cones made from a plexiglas material with a typical refractive index in the order of 1.4 allows to increase the achievable reflective area by a factor of two compared to hollow cones .
since the length of a typical light concentrator for an exit of 1 mm diameter is in the order of 3mm4 mm , weight and light - attenuation , which is dependent on the length of the material crossed , will define a natural limit on the sensor size for which a solid cone is still efficient . for comparison reasons
not only solid ( @xmath9=1.4 ) cones but also intentionally less efficient hollow cones ( @xmath8=1.0 ) are shown .
non - optimum hollow cones are typically used in current cherenkov telescopes , in which the sensitive area of standard photo - detectors ( pmts ) is not a limiting factor . in short : increasing the refractive index , quadratically increases the reflective area of the system .
+ + + another interesting aspect for the final performance of a telescope is the collection of background photons from the diffuse night - sky background . here , eq .
[ eqn : effectivearea ] leads to an interesting conclusion . since the rate of the night - sky background photons per channel scales with the effective reflective area and the solid angle corresponding to the field - of - view of the pixels , the night - sky background rate @xmath57 is proportional to @xmath58 yielding @xmath59 for the range of @xmath60 $ ] , eq.[eqn : focallength2 ] and eq .
[ eqn : f ] yield a correction factor @xmath49 between 1.0625 and 1.25 . with them , eq .
[ eqn : nsb ] can be transformed into @xmath61 , with eq .
[ eqn : systemconstant ] into @xmath62 .
it is immediately apparent that the night - sky background rate scales with the physical entry area of the pixel . assuming only reasonable camera field - of - views between 3 and 13 diameter
, the coefficient @xmath63 is between 9.6 and 14.2 .
this can be interpreted such that the night - sky background rate per pixel can be considered constant within @xmath64 in the first order along the lines of an optimized telescope .
the dependence of @xmath65 on @xmath32 and the camera field - of - view is shown in fig .
[ fig : nsb ] . describing the scale of the night - sky - background rate versus focal ratio @xmath32 and camera field - of - view ( diameter ) for fixed cone material and photo detector size.[fig : nsb],scaledwidth=48.0% ] quantitatively the night - sky background rate @xmath66 is given by @xmath67 where @xmath68 and @xmath69 are the photo - sensor s photo - detection efficiency and the night - sky background intensity , respectively . for silicon photo detectors ( mppc @xcite ) , as used in the fact camera , and the night - sky background at la palma @xcite , @xmath66 is about 150mhz ( much larger than the device dark count rate ) given a reflective area of 10m@xmath0 and a pixel field - of - view of 0.2 .
if a cutoff in the photo - detection efficiency is introduced at 650 nm ( pmt - like behavior ) , this can be further reduced . in general , the night - sky background rate is not a main problem in cherenkov astronomy , as the combined trigger requirements of signal among nearest pixel neighbors and within a short time lead to its very efficient suppression . [ [ general - considerations ] ] general considerations + + + + + + + + + + + + + + + + + + + + + + existing telescopes are usually under - designed , i.e.the photo - detectors are larger than necessary or the light - concentrators do not reach the maximum possible concentration . on the contrary ,
currently so - called silicon photo - multipliers have proven their potential in cherenkov telescopes @xcite . these silicon based
photo - detectors usually have a very limited area , but used in an optimized setup , their effective physical light collection area , i.e.entry of the light concentrator , can be much larger .
if prices of photo - sensors are compared , this has to be taken into account .
not the price per mm@xmath0 physical sensitive area , but the price per cone entrance area or field - of - view , has to be considered . if cheap enough , the signal of several photo - detectors , equipped with individual light concentrators , could even be summed . for a comparison , in terms of effective reflective area , the transmission losses of solid cones and their gain from avoiding fresnel reflection has to be taken into account , as well as the reflection losses of hollow cones .
[ [ a - note - on - timing ] ] a note on timing + + + + + + + + + + + + + + + + for an ideal davies - cotton reflector , the arrival time distribution of an instantaneous parallel beam flash is practically flat .
more precisely , it is linearly decreasing , but looks flat on the small interval .
the number of photons in the arrival time interval @xmath70 $ ] is @xmath71 .
its width @xmath72 is given by @xmath73 , where @xmath74 is the speed of light .
the interval is of the order 1.1ns for a 4 m class reflector ( @xmath75 , up to slightly less than 4.5ns for a 12 m reflector considering @xmath76 .
this short time spread is not a problem for the observation of showers with a small size telescope as it is still small compared to the cherenkov light flash duration . for medium and large size telescopes ,
a slightly different mirror arrangement should be chosen if time spread matters . by a mirror arrangement ,
intermediate between a spherical ( davies - cotton ) and a parabolic design , the time spread can considerably be improved , maintaining the point - spread function almost completely .
while the point - spread function is dominated by the majority of the mirrors , i.e. outermost mirrors , the time spread is dominated by the ones with the largest @xmath77 mirrors , i.e. innermost mirrors .
consequently , moving the innermost mirrors closer to a parabola immediately improves the time - spread while the effect on the point - spread function is rather limited .
ideally , mirrors on a parabola with adapted focal lengths are used , but might be a cost issue . with adapted focal lengths ,
all mirrors are placed at correct focal distance , so that , a similar point - spread function than for the davies - cotton arrangement can be expected . [ [ remarks - about - cta ] ] remarks about cta + + + + + + + + + + + + + + + + + + recent results of fact @xcite show that a reflector in the order of 3.5 m diameter can give already reasonable physics results with current analysis and detector technology . therefore , a 4 m diameter reflector for sst is assumed .
for physics reason , the field - of - view is supposed to be between 9 and 12 ( leading to a reasonable @xmath32 between 1.5 and 1.8 assuming an optical quality of 4 and a tessellation number of 5 ) . requiring a pixel field - of - view in the order of 0.26 , possible solutions could be solid cones with a 36mm@xmath0 g - apd or hollow cones with a 60mm@xmath0 g - apd , see also fig .
[ fig : cta ] .
the manufacturing of 60mm@xmath0 g - apds is under discussion with hamamatsu .
a rough estimate shows that a solid cone for such a device would be about three times longer than for a 9mm@xmath0 as used in fact .
considering the transmission loss of 10% in the fact cones @xcite , which is a very conservative estimate , such cones would have a loss in the order of 35% .
since solid cones avoid the loss from fresnel reflection at the sealing surface and the g - apd surface , the real light loss would only be around 27% assuming that the hollow cone has a reflectivity of 100% which in reality is not true . keeping the pixel field - of - view constant ,
the gain in reflective area corresponds to the refractive index of the cone material squared . in the case of a refractive index of typical poly(methyl methacrylate ) pmma of 1.4 this is a gain of @xmath10100% reflective area , which outperforms the transmission loss significantly . assuming that the manufacturing of a 36mm@xmath0 g - apd would be as easy as of a 60mm@xmath0 g - apd
, one can compare a solution with a 36mm@xmath0 g - apd and a solid cone and a 60mm@xmath0 hollow cone ( assuming perfect reflectivity ) . in this case
, the transmission loss of the solid cone is around 13% compared to 8% fresnel loss for the hollow cone . on the other hand ,
the solution with the hollow cone yields a 15% smaller reflective area ( same pixel field - of - view ) or 27% more pixels ( same reflective area ) .
assuming further that the price of the camera scales with the price of each channels , a reduction of the number of channels by almost 30% reduces the costs for the camera significantly .
since the costs are also dominated by the price for the photo - detectors , and the price of g - apds , in the first order , scales with the sensitive area , it can be estimated that the price for the 36mm@xmath0 g - apds would be almost a factor of two lower than for the larger ones . in figure
[ fig : cta ] possibly solutions for mst and lst designs are shown using g - apds and solid cones . on both cases
it is convenient to sum at least three , or even seven , pixels into one readout channel to keep the ratio @xmath32 low for construction reasons .
, a pixel size four times the point - spread function and solid cones with @xmath78 . for a given @xmath32 and camera field - of - view
, the corresponding reflector diameter and the pixel field - of - view can be read .
[ fig : conclusion],scaledwidth=48.0% ] the davies - cotton design with its simplicity as compared to non validated dual optic systems is assuredly a good option for a wide field - of - view , up to 1012 , high energy cherenkov telescope . with this study , it is possible to scan a wide phase space of the design of cherenkov telescopes or telescope arrays .
this was achieved by a description of the optical performance of davies - cotton reflectors and introduction of the effect of light - concentrators . in particular
, this study provides an analytical description of the optical performance of a tessellated davies - cotton reflector precise enough to enable performance studies without the need for dedicated simulations . by including the effect of the light - collector into the system of equations , the available phase space of design parameters
is reduced to a single parameter , once the photon detector has been chosen and either the pixel field - of - view or the camera field - of - view has been fixed by physics constraints . while the choice of photo sensor is usually defined by the availability on the market ,
constraints on the camera field - of - view are a result of the physics targets .
if these two parameters are fixed , the whole available phase space of possible solution can now be scanned by changing a single input parameter .
it can , for example , be convenient to scan a reasonable range of the focal ratio @xmath32 and derive all other parameters accordingly . from the result ,
the most cost efficient solution , or the one performing best in sense of physics targets can be chosen . for the cherenkov telescope array ,
several design options were presented .
it could be shown that for the small size telescope , considering a camera field - of - view of 9 to 12 , a four meter reflector is enough if 36mm@xmath0 senors are topped with solid cones to achieve a pixel field - of - view in the order or 0.25 to 0.3 at reasonable @xmath32 .
an alternative solution are hollow cones with correspondingly larger sensor area , which is disfavored because of the costs dominated by the sensor . an example plot which easily allows to determine reasonable options from the available phase space
is shown in fig .
[ fig : conclusion ] .
the reflector diameter can easily be re - scaled linearly with the photo sensor size and the refractive index of the cone material . for the medium size and large size telescopes , the most reasonable solution using small sensors would be the summation of three and seven , respectively . equipped with different sum - stages
, these modules could be applied in any telescopes .
larger silicon based sensors , expected soon on the marked , would allow a single - channel / single - sensor solution .
using several small sensors in one channel has the advantage that the application of solid cones is possible in terms of weight and transmission and costs for photo sensors can be kept low due to their at least two times higher compression ratio .
the davies - cotton design @xcite is known to be promising for wide field prime - focus telescopes and was studied earlier analytically @xcite and through simulations @xcite .
however , the parametrizations are moderately accurate and non existing for tessellated reflectors . here , parametrizations are provided , accurate at the percent level up to 12 field - of - view , for the ideal ( non - constructable ) davies - cotton telescope and accurate to a few percent for a realistic davies - cotton telescope with arbitrary tessellation of the reflector . the major issue of the design of a telescope is the reflector and its optical performance .
since design parameters like the field - of - view of a single pixel or the field - of - view of the whole camera are closely related to the reflectors optical performance , it is important to understand the relation between the reflector design and its performance .
unfortunately , neither spherical nor parabolic mirrors can provide both , good optical point spread function for on - axis and inclined rays , at the same time , because the distance between any point on the mirror surface to the focal point does not match the local focal length defined by the local radius of curvature .
furthermore , in the case of a spherical mirror , also the shape of the mirror surface is not ideal compared to a parabolic mirror .
the parabolic shape ensures that parallel rays from infinity are well focused into a single point ( due to the definition of a parabolic surface ) while in the spherical case this is not the case .
that means that in both cases rays hitting the mirror far off its center have their focal point not at the focal plane . in the case of a spherical mirror
they also miss the focal point ( _ aberration _ ) .
consequently , the ideal mirror would be a combination of two properties : a mirror surface which is shaped such that it has the right focal distance at any point , but at the same time any point is correctly oriented , so that focal distance and direction are correct .
since local normal vector and local curvature can not be disentangled such a mirror can only be a theoretical construction .
tessellating the reflector into individual mirrors , this behavior can be approximated , as shown by davies and cotton , if the reflector is built from several spherical mirrors which are placed on a sphere around the focal point . in this case , the reflector can have the correct focal distance locally and , at the same time , the mirror elements can be oriented such that they correctly focus to the focal point .
apart from an improved optical performance for inclined rays , the production of several small and identical mirrors is also much more cost efficient than the production of a single large mirror . since any optical system can always be linearly scaled , in the following a scale factor is chosen such that the reflector diameter corresponds to unity , which is identical to defining @xmath16 with @xmath15 being the focal length and @xmath14 the diameter of the mirror .
the ideal davies - cotton reflector has a non constructable surface .
its shape @xmath82 is spherical with radius of curvature @xmath38 , but its normal vectors are defined to intercept at location @xmath83 .
formally , the surface equation @xmath84 and the surface normal vectors @xmath85 are @xmath86 to have the root - mean - square of the projection of reflected rays on the focal plane along @xmath87 and @xmath88 coinciding with tangential and sagittal resolutions , a rotation around @xmath89 is performed without loss of generality . an incoming ray with vector @xmath90 will therefore be reflected on the surface in the direction @xmath91 and intercept the ( non curved ) focal plane at @xmath92 , generally yielding @xmath93 for the ideal davies - cotton this takes the explicit form @xmath94 it is straightforward to numerically calculate the image centroid @xmath95 and the resolution @xmath96 of such a telescope and to estimate the contribution of various terms to the resolution with a taylor development of @xmath97 and @xmath98 in terms of @xmath87 , @xmath88 and @xmath11 .
the development of terms of the form @xmath99 , with @xmath100 and @xmath101 was found to be sufficient for a percent precision in the resolution parameters .
the tangential and sagittal barycenter in the focal plane ( the image centroid ) for a uniform beam on the primary surface are given by @xmath102 and the corresponding resolution in term of root - mean - square : @xmath103 the upper integration bound in @xmath45 originates from the fact that the optical system was scaled to meet a reflector mirror of @xmath104=1 , hence @xmath45=@xmath105 .
davies - cotton , only leading terms @xmath108 and @xmath109 are retained , at @xmath43 and a maximum off - axis angle of the incoming rays of @xmath110=5 , @xmath111 terms such that @xmath112 and @xmath113 .
as we apply the same conditions to spherical prime - focus design , less terms are present at higher order , i.e.less spherical aberration .
to mirror the result for the ideal davies - cotton , several non leading terms are added giving a consistent picture for both developments as shown in table [ tab : coeff ] . in @xcite and @xcite , order developments for the davies - cotton and sherical mirror , respectively , have been discussed .
the obtained coefficients are repeated here for completeness in table [ tab : coeff ] .
both solutions show up to 20% fractional error @xmath116 , e.g.at @xmath43 . in the above considerations ,
the shadow of a possible detector in the focal plane has been neglected . by changing the lower bound from @xmath118 to @xmath119@xmath120 in expressions [ expr9 ] , obscuration
can easily be quantified .
[ fig : dctelreso - obscur ] shows that obscuration degrades the resolution parameters by about 1.5% at @xmath43 up to about 6% at @xmath121 , 5 off - axis . and
@xmath114 of the exact results , i.e.numerically calculated , with ( dashed ) and without ( solid ) obscuration .
inset shows the ratio of both for @xmath38=\{1,1.4,2 } ( solid , dashed , dotted ) .
, scaledwidth=47.0% ] a realistic implementation of the non - constructable davies - cotton telescope consists in introducing a reflector made of multiple individual spherical mirrors .
the tessellation number is defined as the the number @xmath42 of mirrors in the diagonal . in the limit @xmath42=@xmath122
, it is identical to the ideal davies - cotton design . in practice
, @xmath42 is a number @xmath12330 .
the effective parametrization is presented as a correction to the limited taylor development derived earlier .
the correction is implemented through ray - tracing simulation performed with the mars software ( described in @xcite ) , which do fully reproduce the results obtained earlier in the case of a spherical and ideal davies - cotton reflectors .
although an ideal davies - cotton reflector can not be build in reality , it can be simulated easily .
simulations enable the use of arbitrary tessellation , since analytical solution being not very well suited for this task .
* individual mirrors are hexagonal . for symmetry reasons
each hexagon is rotated by 15 against the x-/y - axis * the mirrors are fixed on a hexagonal grid in the @xmath87/@xmath88-plane with spacing @xmath104 * the diameter of the individual mirrors is defined as @xmath124 , with @xmath8 being the total number of mirrors in the system * their center is located on a sphere around the focal point ( this corresponds to @xmath125 for the ideal davies - cotton ) * the focal length @xmath15 of each mirror is equal to the radius of the sphere , and therefore equal to the focal length of the system * the overall shape of the reflector is also hexagon like * the tessellation number @xmath42 is the number of mirrors in the diagonal * each mirror is oriented to a virtual point in 2@xmath15 ( this corresponds to @xmath126 for the ideal davies - cotton ) * the small effect of obscuration by the focal instrumentation is neglected * the mirror surface is assumed to be ideal empirically , it could be found that introducing a dependence on the tessellation number , the formulas given in [ appendix : a ] for the spherical mirror and the ideal davies - cotton mirror could be unified . for this , a linear dependence at order in @xmath11 , and a quadratically at order @xmath129 for @xmath130 , has been introduced .
additionally , an effective rescaling @xmath131 is needed to reach an accuracy about 5% in the whole simulated range .
the root - mean - square of a tessellated davies - cotton can then be written as @xmath132 \frac{\phi^i}{f^j_{\rm eff}}\right)\ , , \nonumber\\ \delta\eta^2 & = & \frac{1}{2 ^ 4}\left(\sum_{j } \frac{s^\eta_{0,j}}{n^2 f^j_{\rm eff } } + \sum_{i>0,j } \left[\frac{s^\eta_{i , j}-d^\eta_{i , j}}{n}+d^\eta_{i , j}\right ] \frac{\phi^i}{f^j_{\rm eff}}\right)\,.\nonumber\\ & & \label{eq : realdc}\end{aligned}\ ] ] the coefficients @xmath133 and @xmath134 are the ones given in table [ tab : coeff ] for the spherical and the ideal davies - cotton , respectively . as defined by formula [ eq : realdc ] determined from fits to the simulated point - spread function .
each point is the average of all correction factors obtained for the simulated range of @xmath32 .
the error bar denotes the spread .
the two curves are the tangential ( upper curve , solid ) and sagittal ( lower curve , dashed ) , respectively.,scaledwidth=48.0% ] describing the transition from a single spherical mirror to an ideal davies - cotton reflector . the rescaling factor @xmath135
can be interpreted as the deviation of the shape from the ideal case .
its value was determined by minimizing the residual , i.e.@xmath136 , between simulated point - spread function and approximated root - mean - square .
the differences of the sagittal and tangential residual are minimized independently for each @xmath42 .
its value is depicted in fig .
[ fig : correction ] .
the introduction of this scale factor effectively reduced the residual from a maximum of 12% to less than 5% for tessellation numbers
smaller than 40 .
[ fig : residual ] shows the distribution of the residuals for different tessellation numbers .
note that for the case @xmath137 the simulated single mirror is of hexagonal shape while the analytical approximation describes a disc - like mirror . for
cases @xmath138 the properties of the simulated reflector converge to the ideal davies - cotton . while the presented development was calculated for a disc shaped reflector , here , the simulated davies - cotten converge to an ideal hexagon .
consequently , in both cases the rescaling factor is expected to be different from unity . in general
, it is not expected to obtain a perfect match between the analytical approximation and the simulation , because simulations will always take into account effect which can not be easily described analytically , like rays lost between individual mirrors . from eqs .
[ eq : realdc ] it is evident that for rays with small incident angles the point - spread function is dominated by the -order term which decreases fast with high tessellation number . at higher incident angles the point - spread function is dominated by higher order terms which only turn from the spherical to the ideal davies - cotton solution for increasing tessellation numbers .
in general the dominating term for reasonable incident angles is the -order term .
consequently , the point - spread function dramatically improves for @xmath42@xmath1391 but for @xmath42@xmath555 changes become unimportant .
hence , for practical purposes a single mirror and the case of @xmath140 can be excluded while for most practical purposes @xmath141 will already be enough .
it is possible to describe the optical quality of a set of well defined davies - cotton reflectors quite well in a single analytical formula .
even the real davies - cotton might be slightly different , e.g.different mirror or reflector shapes or obscuration by the focal plane instrumentation , this gives a very good estimate of the optical performance .
stamatescu , v. , rowell , g. p. , denman , j. , _ et al .
_ , 2011 , astropart .
phys . , 34 , 886 .
bernlhr , k. 2008 , astropart .
phys . , 30 , 149 .
schliesser , a. , and mirzoyan , r. , 2005 , astropart .
phys . , 24 , 382 .
actis , m. , agnetta , g. , aharonian , f. , _ et al .
_ , 2011 , experimental astronomy , 32 , 193 .
r. winston , j. c. minano , p. benitz , _ nonimaging optics _
, 2004 , elsevier academic press .
braun , i. , _ et al .
_ , 2009 , in proc . of the international cosmic ray conference
huber , b. , braun , i. , _ et al . _ , 2011 , in proc . of the international cosmic ray conference .
anderhub , h. , backes , m. , biland , a. , _ et al .
_ 2011 , nuclear instruments and methods in physics research a , 628 , 107 .
aliu , e. , anderhub , h. , antonelli , l. a. , _ et al . _ , 2009 , astropart .
phys . , 30 , 293 .
anderhub , h. , backes , m. , biland , a. , _ et al .
_ , 2012 , proc . of the international symposium on high - energy gamma - ray astronomy ( _ in press _ )
mppc multi - pixel photon counters , 2008 , hamamatsu .
chris r. benn and sara l. ellison , la palma night - sky brightness .
a. k. konopelko , 1997 , towards a major atmospheric cerenkov detector v , 208 .
davies , j. m. , and cotton , e. s. , 1957 , j. solar energy sci . and eng .
vassiliev v. , fegan s. , brousseau p. , 2007
, astropart .
phys . , 28 , 10 .
schliesser a. , mirzoyan r. , 2005 , astropart .
phys . , 24 , 382 .
bretz , t. and wagner , r. , 2003 , in proc . of the international cosmic ray conference , 5 , 2947 .
bretz , t. , 2005 , high energy gamma - ray astronomy , 745 , 730 .
bretz , t. and dorner , d. , 2008 , american institute of physics conference series , 1085 , 664 . | this paper discusses the construction of high - performance ground - based gamma - ray cherenkov telescopes with a davies - cotton reflector . for the design of such telescopes ,
usually physics constrains the field - of - view , while the photo - sensor size is defined by limited options .
including the effect of light - concentrators in front of the photo sensor , it is demonstrated that these constraints are enough to mutually constrain all other design parameters .
the dependability of the various design parameters naturally arises once a relationship between the value of the point - spread functions at the edge of the field - of - view and the pixel field - of - view is introduced . to be able to include this constraint into a system of equations , an analytical description for the point - spread function of a tessellated davies - cotton reflector
is derived from taylor developments and ray - tracing simulations . including higher order terms
renders the result precise on the percent level .
+ design curves are provided within the typical phase space of cherenkov telescopes .
the impact of all design parameters on the overall design is discussed .
allowing an immediate comparison of several options with identical physics performance allows the determination of the most cost efficient solution .
emphasize is given on the possible application of solid light concentrators with their typically about two times better concentration allowing the use of small photo sensors such as geiger - mode avalanche photo diodes .
this is discussed in more details in the context of possible design options for the cherenkov telescope array . in particular , a solution for a 60mm@xmath0 photo sensor with hollow cone is compared to a 36mm@xmath0 with solid cone .
tev cherenkov astronomy , davies - cotton design parameters , photo - sensors , winston cone |
globular clusters ( gcs ) are old , gravitationally bound stellar systems whose core stellar densities can be extremely high , reaching up to @xmath5 . in such an environment , close encounters and even direct collisions with resulting mergers between the cluster stars are quite common , leading to a variety of dynamically - formed stellar objects like blue stragglers ( bss ) and close binary ( cb ) systems .
cbs are important for our understanding of gc evolution , since the binding energy of a few , very close binaries can rival that of a modest - sized globular cluster .
thus , by transferring their orbital energy to passing single stars , cbs can significantly affect the dynamical evolution of the cluster ( e.g. elson et al .
1987 , hut et al .
1992 , and references therein ) .
this depends critically on the number of cbs .
if there are only a few cbs , long - term interactions dominate the cluster evolution . by contrast
, the presence of many cbs leads to violent interactions , which heat the cluster , and promote its expansion and evaporation .
interacting cbs form a particularly interesting subset of cbs , and can , in principle , also be used as tracers of a cluster s cb population .
the best - known formation channel for interacting cbs in gcs is tidal capture of a red giant or main sequence ( ms ) star by a compact object @xcite .
this scenario was originally proposed to account for the overabundance of low - mass x - ray binaries ( lmxbs ) containing accreting neutron stars ( nss ) in gcs , relative to the galactic field .
however , tidal capture theory predicts a comparable overabundance of interacting cbs with a white dwarf ( wd ) primary , i.e. cataclysmic variables ( cvs ) .
since wds are far more common than nss , we would then also expect many more cvs than lmxbs in gcs .
more recently , it has been realized that 3-body encounters ( e.g. davies & benz 1995 ) and `` ordinary '' evolution of primordial binaries ( e.g. davies 1997 ) may also produce significant populations of cvs in gcs .
cvs produced from primordial binaries are expected mainly in the outskirts of clusters , whereas cvs formed dynamically ( through tidal capture or 3-body processes ) should be found preferentially in the dense cluster cores .
thus , the observation of the relative abundances and distribution of cvs within a cluster s core and halo can tell us about the relative efficiency of these different cv formation scenarios .
however , despite their impact on cluster evolution and their importance for our understanding of cb formation and evolution channels , there have been only a few detections of interacting cbs in gcs during the past decades .
finding and studying these systems has proven to be extremely difficult , since the spatial resolutions and detection limits of most available telescopes are too limited for their detection . only with the improved sensitivity and imaging quality of _ chandra _ and _ hst _
has it become possible to finally detect significant numbers of cvs and other binary systems in gcs ( e.g.grindlay et al .
2001 ; albrow et al . 2001
; edmonds et al . 2003a , 2003b ; knigge et al .
2002 , 2003 ; and references therein ) .
@xmath0-imaging is particularly well - suited for the detection of cvs ( and also bss and young wds ) in gc cores .
this is because these objects are characterized by relatively blue spectral energy distributions and emit significant amounts of radiation in the @xmath0 .
by contrast , the `` ordinary '' stars that make up the bulk of the cluster are too cool to show up in observations at such short wavelengths . as a consequence , crowding is not a severe problem in @xmath0 imaging studies of gcs , even in the dense cluster cores .
this is a significant advantage compared to optical gc surveys . to date , the only deep @xmath0 study of a globular cluster with the principal purpose of identifying cvs ( as well as bss and wds ) has been of 47 tuc @xcite .
there , we found 16 cv candidates ( including 4 variable objects that were previously known or suspected cataclysmics ) , 19 bss ( including 4 variables ) and 17 hot wds .
the population of cv candidates was particularly interesting , since their number was broadly consistent with tidal capture predictions ( but note that most of these candidates have not yet been confirmed ) . here ,
we present a reanalysis of _ hst_-based @xmath0 observations of another gc , namely ngc2808 ( @xmath6 , @xmath7 ) .
this intermediate metallicity gc ( @xmath8 } = -1.36 $ ] , walker 1999 ) lies at a distance of 10.2 kpc and is reddened by @xmath9 mag ( bedin et al .
the cluster has a very dense and compact core , with a core radius of @xmath10 and a tidal radius of @xmath11 @xcite .
the data set that we analyse in the following has already been studied by @xcite , but with a focus on the bright , blue hb stars in the cluster core .
more specifically , @xcite discovered a population of subluminous horizontal branch stars in ngc2808 which might have undergone a late helium flash while descending the wd cooling curve .
by contrast , here we are mainly interested in the dynamically - formed stellar populations ( cvs and bss ) , as well as the young wds in this cluster .
since these are faint objects and normally difficult to detect , we first construct a new , deep colour - magnitude diagram ( cmd ) in which many of these objects show up , and then search for variability amongst our @xmath0 sources ( since this is ubiquitous amongst cvs and also seen in bss located in the instability strip ) . in sect .
[ data ] , we discuss the data and the data reduction . in sect .
[ results ] , we present our far- vs. near - ultraviolet cmd and describe our search for variability among our catalogue stars .
our results are summarized in sect .
[ summary ] .
ngc2808 was observed with the space telescope imaging spectrograph ( stis ) on board the _ hst _ in january / february 2000 .
the original purpose of these observations was to provide geometric distortion corrections for the stis camera
. however , such calibration data are of course also of scientific interest , as already shown by @xcite .
images were taken using the f25qtz and f25cn270 filters , which are located in the far- and near - uv wavebands , with central wavelengths of 1590 ( @xmath0 , f25qtz ) and 2700 ( @xmath12 , f25cn270 ) , respectively .
the plate scale is @xmath13 , resulting in a field of view of @xmath14 for both filters .
the exposure times for the single exposures vary between 480 and 538 sec .
the observations were positioned on the cluster centre ; fig .
[ fuv ] shows the combined images .
the mosaic images were constructed using montage2 @xcite , which registers all the image frames to a single reference frame - incorporating shift , scale , rotation , and distortions determined from positions of the brightest stars .
the resulting @xmath0 and @xmath12 mosaic images show a strange shape that is due to the small offsets and rotation of the single exposures with respect to each other .
the diameter of the mosaics is @xmath15 or @xmath16 .
since the cluster s core radius is @xmath17 , the complete core is covered by our mosaic images . as can be seen immediately , the @xmath12 mosaic is extremely crowded .
crowding , on the other hand , is not a problem in the @xmath0 image .
note that the effective exposure times vary strongly across both mosaics .
this is responsible for the increased background noise in the outskirts of the mosaics , where exposure times are low compared to the central regions ( which have maximum exposure times of 8906 sec [ @xmath12/f25cn270 ] and 8916 sec in the [ @xmath0/f25qtz ] ) . as already mentioned in sect .
[ intro ] , we reinvestigate the same dataset studied by @xcite , but with a different focus . here
, our main goal is to search for faint and/or variable @xmath0 sources ( cvs , bss , wds ) .
we therefore take special care to detect even weak sources in the @xmath0 images , which allows us to construct a very deep cmd . as noted above
, the exposure time is not uniform across the @xmath0 mosaic , which means that the outer parts of the image ( with the lowest exposure times ) have the noisiest background . in order to avoid both a large number of false detections in these regions ( caused by noise peaks ) , and multiple detections in the halos of bright stars
, we created a noise map which we used to `` smooth '' the image .
as a first step , we then used the daofind routine of daophot @xcite running under iraf on the smoothed image to detect @xmath0 sources .
we overplotted the resulting source coordinates on the @xmath0 mosaic and inspected them by eye .
overall , this source finding technique appeared to work well for bright sources ( except for a few multiple detections around bright stars that we deleted by hand ) .
however , it did not work as well for faint sources , especially in the deep , central part of the mosaic .
thus , we added a further 211 sources by hand . in total , our catalogue of @xmath0 sources contains 521 entries .
source detection in the @xmath12 mosaic is a tedious task due to the severe crowding and depends critically on the daofind detection threshold .
if the detection threshold is too high , no faint sources will be detected in the centre of the mosaic ( which has maximum exposure time ) .
the situation can be improved by lowering the detection threshold , but this also leads to false `` detections '' in the noisy outskirts .
( the same is true for the @xmath0 mosaic , but less of a problem , since crowding is much less severe .
we could therefore simply adopt a relatively high threshold in daofind and add the missing sources by hand . )
however , it is important to realize that our goal is not to detect all @xmath12 sources , but to locate @xmath12 counterparts to our @xmath0 sources . in practice
, we therefore chose to work with an intermediate threshold , designed to yield most of the detectable @xmath12 counterparts while keeping the number of false detections at bay .
this is important , because if there are too many false detections , some will be matched erroneously to real @xmath0 sources .
in order to match the @xmath0 and @xmath12 catalogues , we first transformed the @xmath12 coordinates onto the frame of the @xmath0 mosaic .
we then checked for common pairs within a tolerance of 1.5 pixels ; this resulted in 306 matches .
careful inspection of the registered @xmath12 image around unmatched @xmath0 sources revealed an additional 50 @xmath12 counterparts within this tolerance that had not been detected by daofind . increasing
the tolerance to 2 pixels added another 23 matches .
these additional matches are located mainly in the outskirts of the mosaic where our coordinate transformation is slightly less accurate , thus justifying an increased matching tolerance .
all stars listed in @xcite were found in our @xmath0 catalogue ; however , 25 of them did not have @xmath12 matches within 2 pixels .
closer inspection by eye revealed that these stars are either located in the outer regions of the mosaic and had probable @xmath12 counterparts outside the 2 pixel tolerance radius , or they had faint @xmath12 counterparts that are located in the most crowded inner area and were not initially detected by daofind .
we therefore added these missing matches by hand .
a catalogue listing all of our @xmath0 sources is given in table [ catalogue ] .
we include sources without @xmath12 counterparts in table [ catalogue ] , because these are likely to include additional wd and cv candidates .
source may lack a @xmath12 counterpart because ( i ) it is too faint and blue to show up in the @xmath12 mosaic or ( ii ) a counterpart exists , but was not found due to the severe crowding in the @xmath12 image . ] among the sources with @xmath12 counterparts , we have marked all pairs that were matched with the higher tolerance of 2 pixel or were added by hand . in total ,
our catalogue contains 521 @xmath0 sources , including 403 with @xmath12 counterparts .
all 295 sources listed in @xcite are also contained in our catalogue .
aperture photometry was obtained using daophot @xcite .
following @xcite , we chose an aperture radius of 4 pixels and used an annulus between 4 and 10 pixel for sky subtraction .
an exposure map was created that accounts for the different exposure times at different locations in the composite image . for the calibration to the stmag - system we used the aperture corrections derived by @xcite , namely 1.83 and photflam of 1.11e-16 @xmath18 @xmath19 @xmath20 for the f25qtz data , and 1.44 and photflam of 3.92e-17 @xmath18 @xmath19 @xmath20 for the fluxes obtained with f25cn270 .
@xmath19 @xmath20 ( not 3.29e-17 @xmath18 @xmath21 @xmath20 ) ] the results of aperture photometry depend on the coordinates of the source that has to be measured , or , to be more precise , on the `` centering '' .
several re - centering algorithms are available under phot in daophot running under iraf .
@xcite created their coordinate list by centering on each star by eye .
we used their photometry as a point of reference , and hence adopted a gaussian recentering algorithm , since this yielded the best agreement with their measurements .
[ centerings ] displays the magnitude differences between our photometry and that of @xcite , in both @xmath0 and @xmath12 bands .
overall , the agreement is very good , with a few notable exceptions discussed below .
for star no . 516 ( i d . 29 in brown et al . , 2001 )
the difference between our photometry and brown et al.s ( 2001 ) is 0.6 mag in @xmath0 .
this star is located in the outer regions of the @xmath0 mosaic and appears somewhat elongated .
this might be because the object is a blend of two sources , or because the psf is smeared out since the coordinate transformations used to construct our mosaic are least reliable in the outermost regions of the image .
the latter effect would be exacerbated by the asymmetric @xmath0 psf and might cause the brightness to be underestimated when using a small aperture of only 4 pixels .
the situation is similar for star no .
500 ( i d . 13 in brown et al.2001 ) , which is 0.45 mag fainter in our photometry , is also located close to the edges of the mosaic and appears somewhat elongated
. however , a difference greater than 0.1 mag in the @xmath0 occurred only for 17 stars , and a difference greater than 0.15 mag only for 5 stars . for the @xmath12
, we found a magnitude difference of 0.985 mag for star no . 426 ( i d . 189 in @xcite ) .
this seems to be a case where @xcite matched a different @xmath12 source than we did .
though we are quite confident that we correctly matched most of the stars in @xmath0 and @xmath12 , this discrepancy is a useful reminder that occasional mismatches might still occur due to the extreme crowding in the @xmath12 images .
this is especially true for fainter sources , and mismatches probably contribute to the rising @xmath12 differences towards fainter magnitudes .
however , differences greater than 0.15 mag occurred for only 18 stars in the @xmath12 . setting aside these extreme cases , the mean absolute differences between brown et al.s ( 2001 ) magnitudes and ours are @xmath22 mag in @xmath0 and @xmath23 mag in @xmath12 .
thus , in general , there is good agreement between their catalogue and ours .
the remaining differences can be explained by the different methods used to construct the mosaic images , to detect the sources , the different centering algorithms for the photometry and by the severe crowding in the @xmath12 images .
fig . [ cmd ] shows the @xmath1 cmd of ngc2808 s core region .
several distinct stellar populations can be seen , including blue hb stars , bss , wds , but also cv candidates .
our cmd reaches approximately 2 magnitudes deeper than that presented by @xcite .
this is because we took special care in detecting faint sources in the @xmath0 mosaic ( sect .
[ find ] ) .
we also note that we plot all sources independent of their photometric errors , whereas @xcite only included sources satisfying @xmath24 mag and @xmath25 mag in both @xmath0 and @xmath12 .
however , we stress that only 7 of the sources plotted have errors exceeding 0.2 mag in @xmath0 , and all of these have errors in the range 0.2 - 0.3 mag .
we therefore still regard these as significant detections . to aid in the interpretation of the cmd ,
we have also calculated and plotted a set of theoretical tracks . in the following
, we will first briefly describe the construction of the theoretical tracks and then discuss the various stellar populations within our cmd . since main sequence stars and red giants are too cool to show up in our cmd , we only present theoretical tracks for the zero - age horizontal branch ( zahb ) , the zero - age main sequence ( zams ) and the wd cooling curve in ngc2808 . for our zahb track ,
we interpolated on the grid of oxygen - enhanced zahb models provided by @xcite to generate a set of models at the cluster metallicity of @xmath8 } = -1.36 $ ] .
we then used synphot within iraf / stsdas to calculate the @xmath0 and @xmath12 magnitudes of stars on this sequence .
this was achieved by interpolating on the kurucz grid of model stellar atmospheres and folding the resulting synthetic spectra through the response of the appropriate filter+detector combinations .
the zams is included because bss are expected to lie near and slightly to the red of this sequence . in order to generate this track
, we used the fitting formulae of @xcite to estimate the appropriate stellar parameters .
the corresponding @xmath0 and @xmath12 colours were then again estimated from the kurucz model grid within synphot .
finally , our theoretical wd sequence was constructed by interpolating on the @xcite grid of theoretical wd cooling curves , adopting a mean wd mass of 0.55 m@xmath26 .
the models were then again translated to the observational plane by carrying out synthetic photometry with synphot , using a grid of synthetic da wd spectra kindly provided by boris gnsicke ( see gnsicke , beuermann & de martino 1995 ) . for all our synthetic tracks we adopted a distance of 10.2 kpc and a reddening of @xmath9 @xcite .
ngc2808 is one of the most extreme examples among the globular clusters with unusual hb morphology , as first noted by @xcite .
since then this cluster has received considerable attention . in general ,
metallicity is regarded to be the first parameter that influences hb morphology . however ,
several globular clusters show different hb morphologies though their metallicities are similar @xcite .
this led to the suggestion of a second parameter responsible for the unusual behaviour of the hbs in these clusters .
many parameters have been suggested ( e.g. age , mass loss , stellar rotation etc . ) , but a thoroughly convincing explanation is still lacking ( see e.g. rood et al .
1993 , recio - blanco et al . 2004 ) .
ngc2808 shows a bimodal hb and one of the longest blue hb tails , the so - called extreme hb ( ehb ) , with prominent gaps between the red hb ( rhb ) , blue hb ( bhb ) and ehb ( e.g. clement & hazen 1989 , ferraro et al . 1990 ,
byun & lee 1991 , sosin et al .
1997 , walker 1999 , bedin et al .
only three other globular clusters are known that show bimodal hbs as well as gaps along their hbs , namely ngc6229 , ngc6388 , and ngc6441 @xcite .
@xcite discussed various mechanisms that might be responsible for the hb multimodality , none of which represents a satisfactory explanation .
they suggested that a combination of effects might result in such a hb morphology , some of which might be unique to ngc2808 . in our cmd ,
the two distinct clumps of bright stars around @xmath27 mag correspond to the blue ( @xmath28 mag ) and extreme ( @xmath29 mag ) hb stars .
the red hb stars are too cool to show up in the @xmath0 or @xmath12 images . as already reported by @xcite , a gap within the ehb as seen in optical cmds for this cluster
is not evident in the @xmath1 cmd .
instead , the ehb clump is extended towards fainter magnitudes .
@xcite suggested that these subluminous ehb stars ( as well as the ehb gap in the optical cmds ) can be explained by a late helium - core flash that these stars undergo while they descend the wd sequence .
the recent findings of @xcite and @xcite generally support this theory . for a comprehensive analysis of the hb stars in the @xmath1 cmd
, we refer the reader to @xcite .
a trail of stars can be seen in fig [ cmd ] , starting from the bhb clump and reaching down below the theoretical zahb sequence towards the faint red corner of our cmd .
these sources show magnitudes fainter than the zahb , but brighter than the zams .
the location of these objects agrees with the expected location of bss , which are thought to be dynamically - formed objects resulting from a collision or coalescence of two or more ms stars .
they are more massive than normal cluster ms stars and therefore we expect them to be already somewhat evolved
. this explains their location slightly above and to the red of the zams .
we find 61 bs candidates , but caution that this should not be taken as a strict number since the various zones in our cmd partly overlap , and the discrimination between the cv and bs candidates , in particular , is difficult ( see below ) .
[ cmd ] reveals a population of 40 objects that lie near the theoretical cooling curve and are therefore probably hot , young wds . of these
, 22 are brighter than @xmath30 and should be detectable across the full @xmath0 and @xmath12 mosaics .
how many wds should we expect to see ?
following @xcite , we can obtain a simple estimate by scaling from the number of hb stars in the same field of view .
we count @xmath31 bhb and ehb stars in our cmd , but , as noted above , rhb stars are too cool to show up .
we therefore estimate the total number of hb stars by adopting the ratio of ( rhb+bhb+ehb)/(bhb+ehb ) @xmath32 found by @xcite . we therefore expect a total of @xmath33 hb stars in our field of view . given that the lifetime of stars on the hb is approximately @xmath34 yrs ( e.g. dorman 1992 )
, we can predict the number of wds above a given temperature on the cooling curve from the relation @xmath35 where @xmath36 is the wd cooling age at that temperature ( e.g. knigge et al .
wds at @xmath37 mag on our wd sequence have a temperature of @xmath38 k and a corresponding cooling age of @xmath39 yrs .
we therefore predict a population of approximately 16 wds brighter than @xmath30 , which is in reasonable agreement with the observed number of 22 .
this strongly suggests that most of our candidates are indeed wds .
[ cmd ] reveals quite a number of stars that are located between the wd and zams tracks .
this is a region in which we might expect to find cvs .
we estimate that there are @xmath40 sources in this `` cv zone . ''
how does this number compare to theoretical predictions ?
we can attempt to answer this question by scaling the number of tidal capture cvs predicted for 47tuc to ngc2808 . to this end , we use the simplified estimate for the capture rate in globular cluster cores @xcite @xmath41 , where @xmath42 is the central luminosity density and @xmath43 the core radius . based on this , we find that the total number of dynamically - formed cvs in ngc2808 should be quite similar to 47tuc , in which @xcite predicted a population of @xmath44 active cvs formed via tidal capture . for 47tuc
, @xcite found that approximately half of the captures would take place within 1 core radius .
a cv formed outside the core will drift towards the centre due to mass segregation . thus a given active cv should be found in the core today if its age exceeds the relaxation time - scale at the location where it formed .
given our detection limits , we can not hope to detect the very oldest cvs , since these are also the faintest . in order to obtain a rough estimate
, we will assume that we can detect most cvs whose secondaries have not yet been whittled down to brown dwarfs , but none of the latter .
thus our dividing line is between cvs whose orbital periods are still decreasing , and the so - called period bouncers , whose periods are lengthening ( see distefano & rappaport , 1994 , for details ) . based on fig . 2 in @xcite , the oldest cvs we can detect will therefore have ages on the order of @xmath45 yrs .
this is comparable to the relaxation time - scale at the half - mass radius of ngc2808 , so some , but probably not all dynamically - formed active cvs should have had time to reach the core .
given this ambiguity , we assume that between 50% and 100% of active cvs in ngc2808 should be within our field of view .
however , only about half of these are above our adopted detection limit , yielding a final prediction between @xmath46 cvs .
we stress that this is an extremely rough estimate , but it does bracket the observed number of sources in our cv zone .
thus , a significant fraction of these candidates might indeed be cvs .
we caution that other sources might also occupy our cv zone .
we have to consider the possibility that mismatches might have occurred , though special care has been taken in detecting and matching the @xmath0 and @xmath12 sources ( see sect .
[ find ] and [ photometry ] ) .
non - interacting wd / ms binaries without accretion would also lie in the cv zone .
however , according to our zams sequence , a turn - off star with a mass of 0.85 m@xmath26 would have a @xmath12 magnitude of @xmath47 .
these objects would therefore be expected in the cv zone , but close to the wd cooling sequence where we find about 10 objects .
if we do not consider these 10 sources , our number of cv candidates would be rendered to @xmath48 , which is still consistent with our above estimates . in principle
, our cmd may also contain background / foreground sources .
however , we expect field star contamination to be negligible ( see e.g. ratnatunga & bahcall 1985 ) . fig .
[ cumulative ] shows the cumulative radial distributions of the stars that show up in our cmd .
we distinguish between hb stars ( dotted line ) , cv candidates ( dot - short dashed line ) , wd candidates ( long dashed line ) , and bs candidates ( short dashed line ) . the top panel shows the distribution if _ all _ stars of the corresponding population are considered , independent of their magnitude .
however , as pointed out in sect . [ find ] , the exposure time is not uniform across our mosaic , thus also the detection limit varies across the mosaic .
for the outermost regions with the lowest exposure time all sources at least down to @xmath30 mag are detected .
we try to avoid selection effects by adopting this as a limiting magnitude for our entire dataset .
the lower panel in fig .
[ cumulative ] shows the cumulative distributions of the magnitude - selected data .
the most striking differences seem to be between our hbs and wd candidates , which show the least central concentration , and the bss and cv candidates , which appear to be the most centrally concentrated populations .
this is in agreement with @xcite and @xcite who presented deep wide - field photometry and found no radial gradient in the distribution of the ehb stars . in order to compare the cumulative distributions and to establish the statistical significance of the differences between them
, we applied a kolmogorov - smirnov test on the magnitude - selected dataset .
however , we caution that the wd , cv and bs candidates can not clearly be distinguished on the base of our cmd alone , i.e. we might have sources in one sample that actually belong to another one . applying the magnitude selection to our data yields 22 wd , 30 bs , and 16 cv candidates . the hb stars are brighter than our selection criterion , and 209 of them are present in our field of view . the k - s test returns the probability that the maximum difference between the two distributions being compared should be as large as observed , under the null hypothesis that the distributions are drawn from the same parent populations . according to this ,
the distributions of hbs and wds are consistent with one another , with a k - s probability of 64.7 % .
however , the distributions of bs and cv candidates differ significantly from those of the other two populations , with k - s test probabilities of 1.0 % ( cvs vs hbs ) , 0.5 % ( cvs vs wds ) , 0.4 % ( bs vs hbs ) and 1.2 % ( bss vs wds ) . the differences between the cv and bs distributions are not statistically significant , with a k - s probability of 49.3 % ( cvs vs bss ) .
the differences in the observed distributions might reflect the different masses of the various fuv populations .
more specifically , cvs and bss are expected to be considerably more massive than wds and hbs , since cvs are wd / ms binaries and bss are roughly speaking ms stars with masses above the cluster ms turn - off .
the heavy cvs and bss thus sink toward the cluster core and as a result are more centrally concentrated than the lighter wds and hbs .
[ cumulative ] shows that the differences revealed by the k - s test are indeed in the expected direction . on the other hand , cvs and bss
are believed to be objects formed through two- or three - body interactions such as tidal capture , collisions and mergers .
these dynamical interactions take place preferentially in the dense cluster core .
thus , cvs and bss may also be expected to be overabundant in the cluster core because this is their birthplace . in either case , the enhanced central concentration of our cv candidates provides additional evidence that most of these sources are indeed cvs or non - interacting wd / ms binaries ( as opposed to chance superpositions , foreground stars , etc ) , i.e. they either segregated towards the cluster centre , or they formed in the dense core region .
@xcite presented optical photometry for 74 globular clusters , among them ngc2808 .
the clusters were observed with _
hst_/wfpc2 in the @xmath49 and @xmath50 bands , with the pc centred on each cluster s centre .
we searched for optical counterparts to our @xmath0 sources in their dataset , using xyxymatch within iraf .
the transformation between the pc and stis physical coordinate systems is tied to three hb stars as reference objects which could be clearly identified by eye in both the @xmath50 and the @xmath0 images .
following @xcite , we allowed for a maximum difference of 0.8 pc pixels ( corresponding to 1.5 stis pixels ) . in total , we found 97 counterparts to our @xmath0 sources .
out of these 9 are matches with ms stars , red giants or rhb stars .
these stars are too cool to show up in our @xmath1 cmd , and thus these matches have to be explained otherwise : 13156 objects are present in the pc field - of - view in both optical bands , and the complete pc field is covered by our @xmath0 mosaic .
in addition , 390 @xmath0 sources are located inside the area covered by the pc - data . following the same statistical approach as in @xcite
, we would expect up to 13 false matches .
thus , all the matches with ms stars and red giants can be explained statistically as false matches .
the remaining 88 matches are all blue objects , either bhb or ehb or bs stars .
we marked all our @xmath0 sources for which we could find a blue optical counterpart as crosses in our @xmath1 cmd , and added their optical magnitudes and i d from @xcite to our table [ catalogue ] .
as can be seen , one of our cv candidates , our star no . 170 , does have an optical counterpart with @xmath51 mag and @xmath52 mag .
the blue optical colour of the counterpart ( well off the cluster ms ) , and the quality of the @xmath0/optical match ( the @xmath0 and optical positions agree to within 0.4 pixels ) both suggest that the counterpart identification is secure . as a first check for possible variability among our catalogue stars we performed photometry on all single @xmath0 images . for this purpose
we first used the processed images that had been used to create the mosaic , i.e. these images were shifted and rotated to a common logical ( i.e. image pixel ) coordinate system .
the great advantage is that we can use our coordinate list derived from the @xmath0 mosaic as an input to all of these processed images .
the photometry was carried out in the same way as described above in sect .
[ photometry ] .
of course only a fraction of our catalogue stars are present in each of the individual images , and for some stars we could obtain only one measurement .
[ varsigma ] displays a plot of the mean magnitudes versus the corresponding @xmath53 calculated from all individual magnitudes derived for each star .
as can be seen , some of our objects show a considerably higher @xmath53 than the majority . the crosses in fig .
[ varsigma ] denote 28 objects that show a large @xmath53 , compared to their companions with similar brightness , and that we chose to inspect more closely .
we cross - checked our method using median ( instead of mean ) magnitudes and @xmath54 ( instead of @xmath53 ) . in each case
the same stars show an excess @xmath53 ( or @xmath54 ) . as a next step
, we used the non - processed images to get more precise magnitudes for our 28 variable candidates .
this is necessary , because the orientation of the psf on these images is the same with respect to each image and not rotated as for the processed images that were used for creating the mosaic . since these images now do not share a common logical coordinate system
, we had to derive the coordinates for our variable candidates on each single image by using imexam under iraf .
the photometry was carried out in the same way as described in sect .
[ photometry ] .
stars no .
79 , 114 , 175 and 269 of our table [ catalogue ] show very high @xmath53 , however , these stars are located very close to bright stars , which makes it difficult to get accurate coordinates on the single images that have much shorter single exposure times than the mosaic .
consequently , these stars do not stand out as clearly as they do on the @xmath0 mosaic , and sometimes it is not possible to distinguish them from the halo of the nearby bright star .
for 18 of the other stars that we selected on the basis of fig .
[ varsigma ] , the magnitude scatter either disappears when using the single non - processed image photometry , or too few measurements are available for reliable conclusions . however , for none of these 18 sources is a magnitude variation as evident as it is for star no .
358 , see below .
we note that @xcite compared observed and predicted count rates in the @xmath0 f25qtz and found a much larger scatter between observations and predictions than expected from poisson statistics .
only 6 stars out of the 28 candidates have more than 5 measurements and a @xmath53 well beyond the expected photometric scatter also if we take the effect described by @xcite into account and we consider these sources as `` secure '' variable candidates .
the remaining 22 stars are labelled in table [ catalogue ] but are not discussed further here .
ngc2808 was observed on only 4 days ( 18 .
january 2000 and 16 .
february 2000 ) .
this period is sufficient to search for signs of variability , but it is somewhat too short and too interrupted to present reliable lightcurves for our variable candidates
. however , fig .
[ varstars ] shows the magnitudes versus time for the 6 sources that remain as good variable candidates after our analysis of the single , non - processed images , i.e. these stars still show a brightness variation that is well outside the expected magnitude scatter .
the 6 variable candidates are overplotted as diamonds in fig .
star no .
222 lies in the cv zone at @xmath55 mag , making it a good cv candidate . star no . 397 is located close to the wd region in our cmd , but could certainly be a cv as well . as mentioned above , the wd and cv zones are likely to overlap and can not clearly be distinguished on the base of our cmd alone .
the lightcurve for star no .
124 in fig .
[ varstars ] seems to suggest that this source might be an eclipsing binary , with the beginning of an eclipse indicated at the end of the first observing period .
this star is located on our theoretical zams .
it is very likely that this star is a bs star , and some bs are known to be binaries ( see livio 1993 , and references therein ) .
stars no .
162 and 76 are clearly located in the bs zone of our cmd .
their lightcurves , especially for star no . 162 , seem to indicated that these stars are pulsating variables ( for other examples of oscillating bss , see gilliland et al .
1998 , and references therein ) .
star no .
358 is presented in the top panel in fig .
[ varstars ] .
this is the brightest of our variable candidates , located at @xmath56 mag and @xmath57 mag and thus above the bs zone .
it has a @xmath58 mag in fig .
[ varsigma ] , and its brightness drops about 4 magnitudes from the january to the february observations .
this can be seen by eye on the @xmath0 images , as illustrated in fig .
the image on the left hand - side of this figure represents a close up of image o60q02f6q .
star no .
358 clearly sticks out as one of the brighter objects ( @xmath59 mag ) . the image on the right hand - side of fig .
[ no70 ] is a close up of exposure o60q52kxq , centred around the same region .
as can be seen , star no .
358 has practically vanished .
we found that the coordinates of this star agree within @xmath60 in @xmath61 with the rr lyrae v22 of @xcite .
@xcite pointed out that the amplitude of rr lyrae can be as high as 4 mag in the @xmath0 .
also , the lightcurves of rr lyrae are known to be asymmetric with a period of 0.1 to 0.3 days .
it seems that we observe a brightness peak in the first observation period and a minimum in the second one .
recently , @xcite found @xmath0 variations of about 4.9 mag within the 0.56432 d period of their rr lyrae star , which is of the same order as the @xmath0 amplitude we found for our star no . 358 .
large @xmath0 amplitudes are to be expected since the amplitudes of the pulsations increase towards the far - ultraviolet @xcite .
the uv photospheric flux is extremely sensitive to @xmath62 in the temperature range of rr lyrae stars ( @xmath63 k ) .
we have reanalyzed far - uv stis _ hst _ observations of the core region of the globular cluster ngc2808 .
these data were first analyzed by @xcite with an emphasis on the bright bhb and ehb stars , whereas our focus has been on the population of fainter @xmath0 sources .
taking special care in detecting the faint @xmath0 sources , we were able to present a @xmath1 cmd that is @xmath64 mag deeper than the one presented by @xcite .
the overall agreement between their photometry and ours is good .
various stellar populations show up in our cmd , including the bhb and ehb stars for which this cluster is well known .
rhb and ms stars are too cool to show up .
however , approximately 60 bs candidates are present in our cmd .
about 40 sources are located close to our theoretical wd cooling curve and are probably hot , young wds .
we estimated the number of wds by scaling from the number of hb stars that are present in our field of view , and found that approximately 16 wds can be expected at @xmath65 mag .
this is in reasonable agreement with the observed number of 22 at a completeness limit of 21 mag .
our cmd reveals a number of stars that are located in the cv region between the wd cooling curve and the zams .
approximately 60 sources can be found in this region , but we caution that due to the photometric scatter the discrimination between the individual zones can be difficult . using a simplified estimate for the capture rate in globular cluster cores @xcite , we found that the number of dynamically - formed cvs in the core of ngc2808 is comparable to 47tuc , and broadly consistent with the number of candidates we detected .
this suggests that many of our candidates might indeed be cvs .
the cumulative radial distributions of the individual stellar populations indicate that our bs and cv candidates are more centrally concentrated than hbs and the wd candidates .
this may be a result of mass segregation , but might also reflect the fact that cvs and bs are mostly dynamically - formed populations that were preferentially born in the dense cluster core .
we compared our @xmath0 data with optical _ hst_/wfpc2 data presented by @xcite and found 88 optical counterparts for our hb and bs stars and for one cv candidate ( our star no . 170 ) .
we searched for variability among our catalogue stars and found six good variable candidates .
these stars are all located at @xmath66 mag and distributed over the entire colour range of our cmd .
two variable candidates , stars no . 222 and 397 ,
lie in the cv zone of our cmd and thus are excellent cv candidates .
stars no . 124 ,
162 and 76 are all located in the bs zone of our cmd .
the lightcurve of star no .
124 suggests that this source may be an eclipsing binary , while the other two stars might be pulsating variables .
the most striking among our variable stars is our star no .
this is the brightest variable in our sample and shows a considerable drop off about 4 mag during the observation period .
we identified this star as the rr lyrae v22 in @xcite .
rr lyrae can show large @xmath0 amplitude variations of even up to @xmath67 mag ( see downes et al .
2004 , wheatley et al .
2004 ) .
albrow , m. d. , gilliland , r. l. , brown , t. m. et al .
2001 , apj , 559 , 1060 bedin , l. r. , piotto , g. , zoccali , m. et al .
2000 , a&a , 363 , 159 brown , t.m . ,
sweigart , a.v . ,
lanz , t. , landsman , w.b . , & hubeny , i. 2001 , apj , 562 , 368 byun , y. , & lee , y. , 1991 , in asp conf .
ser . 13 : the formation and evolution of star clusters , p. 243 catelan , m. , boressova , j. , sweigart , a. v. , & spassova , n. 1998 , apj , 494 , 265 clement , c. m. , & hazen , m. l. 1989 , aj , 97 , 414 corwin , t. m. , catelan , m. , borissova , j. , & smith , h. a. 2004 , a&a , 421 , 667 davies , m. b. , & benz , w. 1995 , mnras , 276 , 887 davies , m. b. 1997 , mnras , 288 , 117 distefano , r. , & rappaport , s. 1994 , apj , 423 , 274 dorman , b. 1992 , apjs , 81 , 221 downes , r. a. , margon , b. , homer , l. , & anderson , s. f. 2004 , aj , 128 , 2288 edmonds , p. d. , gilliland , r. l. , heinke , c. o. , & grindlay , j. e. 2003a , apj , 596 , 1177 edmonds , p. d. , gilliland , r. l. , heinke , c. o. , & grindlay , j. e. 2003b , apj , 596 , 1197 elson r , hut p. , & inagaki s. 1987 , , 25 , 565 fabian , a. c. , pringle , j. e. , & rees , m. j. 1975 , mnras , 172 , 15p ferraro , f. r. , fusi pecci , f. , buonanno , r. , & alcaino , g. 1990 , a&as , 84 , 59 gnsicke , b. , beuermann k. , & demartino , d. 1995 , a&a , 303 , 127 grindlay , j. e. , heinke c. , edmonds , p. d. , & murray , s. s. 2001 , science , 292 , 2290 gilliland , r. l. , bono , g. , edmonds , p. d. et al . 1998 , apj , 507 , 818 harris , w. e. 1974 , apj , 192 , l161 harris , w.e . 1996 ,
aj , 112 , 1487 heinke , c. o. , gindlay , j. e. , lugger , p. m. et al .
2003 , apj , 598 , 501 hut , p. , mcmillan , s. , goodman , j. , mateo , m. , phinney , e. s. et al .
1992 , pasp , 104 , 981 knigge , c. , zurek , d. r. , shara , m. m. , & long , k. s. 2002 , apj , 579 , 752 knigge , c. , zurek , d. r. , shara , m. m. , long , k. s. , & gilliland , r. l. 2003 , apj , 599 , 1320 livio , m. 1993 , asp conf . ser . 53 : blue stragglers , p. 3
moehler , s. , sweigart , a. v. , landsman , w. b. , hammer , n. j. & dreizler , s. 2004 , a&a , 415 , 313 momany , y. , bedin , l. r. , cassisi , s. et al .
2004 , a&a , 420 , 605 piotto , g. , king , i. r. , djorgovski , s. g. et al .
2002 , a&a , 391 , 945 proffitt , c. r. , brown , t. m. , mobasher , b. , & davies , j. 2003 , instrument science report stis 2003 - 01 ratnatunga , k. u. , & bahcall , j. n. 1985 , apjs , 59 , 63 recio - blanco , a. , piotto , g. , apapricio , a. , & renzini , a. 2004 , a&a , 417 , 597 rich , r. m. , sosin , c. , djorgovski , s. g. et al .
1997 , apjl , 484 , 25 rood , r. t. , crocker , d. a. , fusi pecci , f. et al .
1993 , asp conf .
ser . 48 : the globular cluster - galaxy connection , p. 218 sosin , c. , dorman , b. , djorgovski , s. , g. et al . 1997 ,
apjl , 480 , 35 stetson , p. b. 1991 , in 3rd eso / st - ecf garching - data analysis workshop , eds .
grosbl p. j. , warmels r. h. , p. 187
stetson , p. 1994 , pasp , 106 , 250 tout , c. a. , pols , o. r. , eggleton , p. p. , & han , z. 1996 , mnras , 281 , 257 walker , t. 1999 , aj , 118 , 432 wheatley , j. m. , welsh , b. y. , siegmund , o. h. w. et al .
2004 , apj , 619 , l123 wood , m. a. 1995 , lecture notes in physics , 443 , 41 rcccccccccccccl no . & alpha & delta & @xmath0 & @xmath68 & @xmath12 & @xmath69 & @xmath70 & @xmath71 & radius & @xmath72 & @xmath73 & @xmath74 & @xmath75 & comment + & [ @xmath76 & [ @xmath77 & [ mag ] & [ mag ] & [ mag ] & [ mag ] & [ mag ] & [ mag ] & [ @xmath78 & & & [ mag ] & [ mag ] & + 1 & 9 12 5.454 & -64 51 26.761 & 20.808 & 0.170 & 22.187 & 0.229 & - & - & 26.51 & - & - & - & - & a + 2 & 9 12 5.313 & -64 51 26.404 & 20.074 & 0.121 & - & - & - & - & 26.28 & - & - & - & - & + 3 & 9 12 4.414 & -64 51 25.250 & 21.729 & 0.260 & 19.842 & 0.062 & - & - & 24.54 & - & - & - & - & + 4 & 9 12 5.149 & -64 51 28.025 & 17.131 & 0.023 & 18.415 & 0.022 & 0.029 & -0.025 & 24.34 & 253 & - & - & - & pv + 5 & 9 12 5.969 & -64 51 30.908 & 22.210 & 0.229 & 23.169 & 0.379 & - & - & 25.69 & - & - & - & - & pv + 6 & 9 12 4.898 & -64 51 27.503 & 21.911 & 0.198 & - & - & - & - & 23.92 & - & - & - & - & + 7 & 9 12 6.306 & -64 51 32.282 & 18.579 & 0.044 & 20.252 & 0.073 & 0.261 & -0.152 & 26.50 & 289 & - & - & - & n + 8 & 9 12 5.210 & -64 51 28.903 & 16.468 & 0.017 & 16.938 & 0.011 & 0.112 & 0.022 & 23.88 & 256 & - & - & - & pv , e + 9 & 9 12 6.131 & -64 51 32.254 & 16.448 & 0.017 & 17.148 & 0.012 & 0.032 & -0.088 & 25.63 & 287 & - & - & - & e + 10 & 9 12 5.059 & -64 51 28.739 & 18.869 & 0.050 & 18.425 & 0.022 & -0.019 & -0.065 & 23.43 & 252 & - & - & - & + 11 & 9 12 4.342 & -64 51 26.459 & 16.422 & 0.023 & 17.003 & 0.016 & 0.108 & -0.033 & 23.26 & 219 & - & - & - & n , e + 12 & 9 12 6.749 & -64 51 34.589 & 16.458 & 0.023 & - & - & - & - & 27.63 & - & - & - & - & e + 13 & 9 12 4.119 & -64 51 25.910 & 21.532 & 0.275 & - & - & - & - & 23.26 & - & - & - & - & + 14 & 9 12 5.746 & -64 51 31.458 & 20.506 & 0.105 & 19.294 & 0.033 & - & - & 24.27 & - & - & - & - & pv + 15 & 9 12 5.530 & -64 51 30.936 & 19.978 & 0.082 & 17.486 & 0.014 & - & - & 23.65 & - & - & - & - & pv + 16 & 9 12 3.889 & -64 51 25.800 & 19.452 & 0.091 & - & - & - & - & 22.96 & - & - & - & - & + 17 & 9 12 3.460 & -64 51 24.399 & 16.507 & 0.017 & 17.141 & 0.012 & -0.057 & -0.041 & 23.80 & 153 & - & - & - & n ,
e + 18 & 9 12 3.147 & -64 51 23.712 & 16.475 & 0.023 & 17.072 & 0.016 & -0.005 & -0.032 & 24.31 & 128 & - & - & - & a + 19 & 9 12 3.336 & -64 51 24.893 & 16.422 & 0.016 & 17.201 & 0.012 & 0.018 & -0.071 & 23.20 & 145 & - & - & - & n , e + 20 & 9 12 5.822 & -64 51 33.490 & 16.325 & 0.016 & 16.810 & 0.010 & -0.025 & -0.020 & 23.29 & 280 & - & - & - & + | we present a reanalysis of far - ultraviolet ( @xmath0 ) observations of the globular cluster ngc2808 obtained with the _ hubble space telescope_. these data were first analyzed by brown and coworkers , with an emphasis on the bright , blue horizontal branch ( hb ) stars in this cluster . here
, our focus is on the population of fainter @xmath0 sources , which include white dwarfs ( wds ) , blue stragglers ( bss ) and cataclysmic variables ( cvs ) .
we have therefore constructed the deepest @xmath1 colour - magnitude diagram of ngc2808 and also searched for variability among our @xmath0 sources .
overall , we have found @xmath2 wd , @xmath3 bs and @xmath3 cv candidates ; three of the bss and two of the cv candidates are variable . we have also recovered a known rr lyrae star in the core of ngc2808 , which exhibits massive ( @xmath4 4 mag ) @xmath0 variability .
we have investigated the radial distribution and found that our cv and bs candidates are more centrally concentrated than the hbs and wd candidates .
this might be an effect of mass segregation , but could as well be due to the preferential formation of such dynamically - formed objects in the dense cluster core .
for one of our cv candidates we found a counterpart in wfpc2 optical data published by piotto and coworkers . |
during the last decade , several significant model - free formulas , describing the asymptotic behavior of the implied volatility at extreme strikes , were found .
we only mention here r. lee s moment formulas ( see @xcite ) , tail - wing formulas due to s. benaim and p. friz ( see @xcite , see also @xcite ) , asymptotic formulas with error estimates established by the author ( see @xcite ) , andhigher order formulas found by k. gao and r. lee ( see @xcite ) .
we refer the interested reader to the book @xcite by the author for more information .
the present work was inspired by the paper @xcite of s. de marco , c. hillairet , and a. jacquier .
the authors of @xcite obtained interesting results concerning the asymptotic behavior of the implied volatility at small strikes in the case where the asset price distribution has an atom at zero ( see theorem 3.7 in @xcite ) .
special examples of such models are the constant elasticity of variance model , jump - to - default models , and stochastic models described by processes stopped at the first hitting time of zero ( more information can be found in @xcite ) .
it is not hard to see that the right - wing behavior of the implied volatility in models with and without atoms is similar .
therefore , general model - free asymptotic formulas for the implied volatility at large strikes , discussed in chapter 9 of @xcite , can be used in stochastic asset price models with atoms .
however , the left - wing behavior of the implied volatility in models with and without atoms is qualitatively different . this fact was noticed and explored in @xcite .
it was also shown in @xcite that the general formula formulated in corollary 9.31 in @xcite , which describes the left - wing behavior of the implied volatility in terms of the put pricing function , holds for asset price models with atoms ( see formula ( [ e : a ] ) below ) .
for such models , the above - mentioned formula provides only the leading term in the asymptotic expansion of the implied volatility and an error estimate .
the authors of @xcite found a sharper asymptotic formula , characterizing the left - wing behavior of the implied volatility in models with atoms ( see formula ( [ e : one ] ) below ) .
note that the impact of an atom at zero on the left - wing asymptotics of the implied volatility was not taken into account in section 9.9 of @xcite .
this omission led to an incorrect description of the asymptotic behavior of the implied volatility at small strikes in the cev model ( see formula ( 11.22 ) in theorem 11.5 in @xcite ) .
only the absolutely continuous part of the distribution of the asset price was taken into account in formula ( 11.22 ) mentioned above , while the influence of the atom at zero was ignored . in this paper
, we establish new asymptotic formulas for the implied volatility at small strikes in models with atoms ( see formulas ( [ e : imvol2 ] ) , ( [ e : imvol222 ] ) , and ( [ e : imvol22 ] ) below ) .
these formulas contain three explicit terms in the asymptotic expansion of the implied volatility and an error estimate .
note that the asymptotic formula found in @xcite contains two terms and only an incomplete information about the third term is provided .
moreover , there is a qualitative difference between the new formulas and the de marco - hillairet - jacquer formula . in the new formulas , we use the inverse function of a strike - dependent function , while the inverse function of the cumulative standard normal distribution function is employed in @xcite .
it is shown numerically in section [ s : num ] of the present paper that formula ( [ e : imvol22 ] ) provides a significantly better approximation to the left wing of the implied volatility in the constant elasticity of variance model than the de marco - hillairet - jacquier formula .
our next goal is to introduce several known objects , which will be used in the rest of the paper , and then formulate our main results .
the asset price will be modeled by a non - negative martingale @xmath0 defined on a filtered probability space @xmath1 .
the initial condition for the process @xmath0 is denoted by @xmath2 , and it is assumed that @xmath2 is a positive number .
it is also assumed that the interest rate is equal to zero . in the sequel
, the symbols @xmath3 and @xmath4 stand for the call and put pricing functions , associated with the price process @xmath0 .
these functions are defined as follows : @xmath5\quad\mbox{and}\quad p(t , k)=\mathbb{e}\left[(k - x_t)^{+}\right].\ ] ] in the previous formulas , @xmath6 is the strike price , and @xmath7 is the maturity . the implied volatility @xmath8 is the function , satisfying the following condition : @xmath9 the expression on the left - hand side of ( [ e : iv ] ) is the call pricing function in the black - scholes model with the volatility parameter equal to @xmath10 . the function @xmath11 is defined by @xmath12 where @xmath13 is the standard normal cumulative distribution function , that is , the function @xmath14 the functions @xmath15 and @xmath16 in ( [ e : cumul ] ) are defined by @xmath17 and @xmath18 [ r : r ] it will be assumed throughout the paper that @xmath19 for all @xmath20 and @xmath21 .
moreover , we assume that @xmath22 for all @xmath20 and @xmath23 .
the first of the previous restrictions guarantees that @xmath24 exists for all @xmath21 , while under the second restriction , the implied volatility @xmath24 exists for all @xmath25 ( more details can be found in section 9.1 of @xcite ) .
[ r:11 ] the maturity @xmath20 will be fixed throughout the paper . to simplify notation
, we will suppress the symbol @xmath7 in the function @xmath3 and in similar functions .
[ r : navigate ] in the proofs of the results obtained in the present paper , we often assume that @xmath26 .
it is easy to understand why this assumption does not restrict the generality .
indeed , let us define a new stochastic process by @xmath27 , and denote the corresponding call pricing function and the implied volatility by @xmath28 and @xmath29 , respectively . then
, it is not hard to see that @xmath30 .
moreover , the same formula holds for the black - sholes call pricing function , and therefore @xmath31 since the process @xmath32 has @xmath33 as its initial condition , formula ( [ e : nownot ] ) allows us to navigate between asymptotic formulas for the implied volatility under the restriction @xmath26 and similar formulas in the general case .
let @xmath20 , @xmath26 , and set @xmath34 .
if for every @xmath20 , we have @xmath35 , then the function @xmath36 defined by @xmath37 is a call pricing function ( see @xcite ) .
the function @xmath36 plays the role of a link between the left - wing and the right - wing asymptotics of the implied volatility ( see @xcite ) . now , suppose @xmath38 for some @xmath20 .
let us fix such a maturity @xmath7 , and consider @xmath3 , @xmath4 , and @xmath8 as functions of the strike price @xmath6 .
note that for models with atoms , the function @xmath36 , given by ( [ e : g ] ) , is not a call pricing function .
indeed , the function @xmath36 does not satisfy the condition @xmath39 as @xmath40 .
however , the function @xmath36 has many features of a call pricing function .
for example , it is not hard to see that the black - scholes implied volatility @xmath41 exists for all @xmath42 ( see remark [ r : r ] ) , and , in addition , @xmath43 for all @xmath6 with @xmath44 .
we also have @xmath45 where @xmath46 is a positive function such that @xmath47 the proof of ( [ e : r2 ] ) and ( [ e : m ] ) is simple .
indeed , it follows from the definition of the put pricing function that @xmath48,\ ] ] where @xmath49 is the distribution of @xmath50 on the open half - line @xmath51 .
hence , @xmath52 now it is clear that ( [ e : lim ] ) implies ( [ e : r2 ] ) .
finally , the proof of the equality in ( [ e : r1 ] ) for asset price models with atoms is the same as the proof of lemma 9.23 in @xcite .
[ r : distd ] we denote by @xmath53 and @xmath54 the cumulative distribution functions of the random variable @xmath50 on @xmath55 and @xmath51 , respectively .
these functions are given by @xmath56 it is clear that @xmath57 @xmath58 we have already mentioned the asymptotic formula in corollary 9.31 in @xcite .
this formula is as follows : @xmath59 \nonumber \\ &
\quad+o\left(\left(\log\frac{k}{p(k)}\right)^{-\frac{1}{2}}\log\log\frac{k}{p(k)}\right ) \label{e : a}\end{aligned}\ ] ] as @xmath60 . using the mean value theorem and the formula @xmath61 for small values of @xmath6 ( the previous formula follows from ( [ e : g ] ) and ( [ e : r2 ] ) ) ,
we obtain @xmath62 as @xmath60 .
this means that in the presence of atoms , the general formula given in ( [ e : a ] ) provides only the leading term in the asymptotic expansion of the implied volatility near zero .
the expression for the leading term given in ( [ e : zero ] ) can also be predicted from lee s moment formula ( see @xcite ) .
indeed , for models with atoms , all the moments of negative order of the distribution of the asset price @xmath50 explode .
we will next formulate the main result of @xcite , adapting it to our notation .
suppose there exists @xmath63 such that @xmath64 as @xmath65 .
then @xmath66 as @xmath60 . in ( [ e : one ] )
, the symbol @xmath67 stands for the inverse function of the standard normal cumulative distribution function @xmath13 . in addition
, the function @xmath68 in ( [ e : one ] ) satisfies a special estimate of order @xmath69 as @xmath60 ( see theorem 3.7 in @xcite for more details ) . note that the term @xmath70 in ( [ e : one ] ) is not really the third term in the asymptotic expansion of @xmath8 , because of an interplay between the expression in ( [ e : inter ] ) and the function @xmath71 as @xmath60 . in theorem [ t : corrf ] and corollary
[ c : sims ] below , we provide asymptotic formulas for the implied volatility @xmath8 at small strikes with three terms and error estimates of order @xmath72 as @xmath60 .
the main novelty in our approach is that instead of the function @xmath67 used in formula ( [ e : one ] ) , we employ a family of strike - dependent inverse functions @xmath73 , @xmath74 , where @xmath75 it is easy to see that for every @xmath74 , the function @xmath76 is strictly increasing on the interval @xmath77 ( differentiate ! ) , and moreover @xmath78 it follows from the reasoning above that the inverse function @xmath73 exists for all @xmath79 with @xmath80 is strictly increasing on the interval @xmath81 , and maps this interval onto the interval @xmath77 .
[ r : notethat ] note that @xmath82 exists for all @xmath79 such that @xmath83 and therefore , @xmath82 exists for all @xmath79 with @xmath84 . on the other hand ,
if @xmath85 , then we have to assume that condition ( [ e : mainr ] ) holds .
according to remark [ r : notethat ] , @xmath86 is defined for all @xmath74 provided that @xmath87 .
in addition , if @xmath88 , then @xmath86 is defined under the following restriction : @xmath89 note that , given @xmath90 , there exists @xmath91 such that ( [ e : acco1 ] ) holds for all @xmath92 and moreover @xmath93 .
therefore , for all @xmath92 , we can solve the equations @xmath94 and @xmath95 by inverting the function @xmath76 .
the next lemma is simple , and we omit the proof .
[ l : analysis ] let @xmath74 , and suppose @xmath79 satisfies ( [ e : accord ] )
. then the inequality @xmath96 holds if and only if @xmath97 the next two statements are the main results of the present paper . [
t : corrf ] let @xmath98 .
then the following asymptotic formula holds for the implied volatility @xmath24 in the asset price models such as above : @xmath99 ^ 2 \left(\log\frac{x_0}{k}\right)^{-\frac{1}{2 } } \nonumber \\ & \quad+o\left(\left(\log\frac{x_0}{k}\right)^{-\frac{3}{2}}\right ) \label{e : imvol2}\end{aligned}\ ] ] as @xmath60 .
[ c : simson ] let @xmath98 , and suppose the random variable @xmath50 is such that @xmath100 as @xmath60 .
then @xmath101 ^ 2 \left(\log\frac{x_0}{k}\right)^{-\frac{1}{2 } } \nonumber \\ & \quad+o\left(\left(\log\frac{x_0}{k}\right)^{-\frac{3}{2}}\right ) \label{e : imvol222}\end{aligned}\ ] ] as @xmath60 .
[ c : sims ] let @xmath98 , and suppose the random variable @xmath50 satisfies condition ( [ e : ost ] ) .
then @xmath102 ^ 2 \left(\log\frac{x_0}{k}\right)^{-\frac{1}{2 } } \nonumber \\ & \quad+o\left(\left(\log\frac{x_0}{k}\right)^{-\frac{3}{2}}\right ) \label{e : imvol22}\end{aligned}\ ] ] as @xmath60 .
it is not hard to see that corollary [ c : sims ] follows from theorem [ t : corrf ] , remark [ r : distd ] , formulas ( [ e : r2 ] ) and ( [ e : lim ] ) , and the mean value theorem .
[ r : vdrug ] it is interesting to notice that in formulas ( [ e : imvol2 ] ) and ( [ e : imvol22 ] ) , the term of order @xmath103 is absent .
this happens because of certain cancellations , which occur when we combine formulas ( [ e : ex1 ] ) and ( [ e : imvol3 ] ) in the proof of theorem [ t : corrf ] .
[ r : between ] the smile approximations provided in formulas ( [ e : imvol2 ] ) and ( [ e : imvol222 ] ) take into account the distribution @xmath104 of the asset price @xmath50 , while the approximations in formula ( [ e : imvol22 ] ) and in the de marco - hillairet - jacquier formula ( see ( [ e : one ] ) ) use only the mass @xmath90 of the atom at zero .
[ r : vdrugs ] comparing the de marco - hillairet - jacquier formula ( [ e : one ] ) and our formula ( [ e : imvol22 ] ) , one notices two main differences .
first of all , formula ( [ e : imvol22 ] ) contains the expression @xmath105 instead of the expression @xmath106 appearing in formula ( [ e : one ] ) . on the other hand ,
the function @xmath68 in ( [ e : one ] ) satisfies @xmath107 as @xmath60 , while the error term in ( [ e : imvol22 ] ) is @xmath108 as @xmath60
. we will prove theorem [ t : corrf ] in section [ s : next ] , while in section [ s : appl ] we will derive the de marco - hillairet - jacquier formula from our theorem [ t : corrf ] .
in addition , in section [ s : appl ] we give estimates for the difference @xmath109 at small strikes
. section [ s : cevm ] deals with the left - wing behavior of the implied volatility in the cev model . finally , in the last section of the paper ( section [ s
: num ] ) , we compare the performance of two formulas , providing approximations to the implied volatility at small strikes in the cev model : the de marco - hillairet - jacquier formula and the formula in corollary [ c : sims ] in the present paper .
we have already mentioned that it suffices to prove the theorem in the case where @xmath26 . for every small number @xmath63 , set @xmath110 the following assertion provides two - sided estimates for the implied volatility .
[ t : ochen ] let @xmath63
. then there exists @xmath111 such that @xmath112 for all @xmath113 . in ( [ e : wildi3 ] ) , the functions @xmath114 and @xmath115 are defined as follows : @xmath116 ^ 2 \nonumber \\ &
\quad+(u_k)^{-1}(g(k))\sqrt{\left[(u_k)^{-1}(g(k))\right]^2 + 2\log k } \label{e : fina7}\end{aligned}\ ] ] and @xmath117 ^ 2 \nonumber \\ & \quad+(u_k)^{-1}(\widetilde{g}_{\varepsilon}(k ) ) \sqrt{\left[(u_k)^{-1}(\widetilde{g}_{\varepsilon}(k))\right]^2 + 2\log k}. \label{e : final2}\end{aligned}\ ] ] _ proof of theorem [ t : ochen ] .
our first goal is to find two functions @xmath118 and @xmath119 satisfying the following conditions : @xmath120 for sufficiently large values of @xmath6 .
the inequalities in ( [ e : fina2 ] ) will allow us to estimate the implied volatility @xmath41 for large strikes , and hence to characterize the left - wing behavior of the implied volatility @xmath8 .
_ let @xmath121 be a real function growing slower than @xmath122 .
the function @xmath121 will be chosen later .
put @xmath123 then we have @xmath124 and @xmath125 therefore , @xmath126 our next goal is to estimate the last term in ( [ e : ro1 ] ) .
we will use the following known inequalities : @xmath127e^{-\frac{x^2}{2 } } \le\frac{1}{\sqrt{2\pi}}\int_x^{\infty}e^{-\frac{y^2}{2}}dy \le\frac{1}{\sqrt{2\pi}}\frac{1}{x}e^{-\frac{x^2}{2}}. \label{e : sestim1}\ ] ] the estimates in ( [ e : sestim1 ] ) follow from stronger inequalities formulated in @xcite , 7.1.13 . taking into account ( [ e : sestim1 ] ) , we see that @xmath128 and @xmath129 } \nonumber \\ & \quad\exp\left\{-\frac{\varphi(k)^2}{4(\log k+\varphi(k))}\right\}. \label{e : sestim3}\end{aligned}\ ] ] let us next suppose that the function @xmath121 has the following form : @xmath130 where @xmath68 is such that @xmath131 in ( [ e : constant ] ) , @xmath132 is a real number
. then we have @xmath133 \nonumber \\ & \le\frac{3\phi(k)^2}{16(\log k)^{\frac{3}{2 } } } , \label{e : rur}\end{aligned}\ ] ] for all @xmath134 .
it follows from ( [ e : sestim2 ] ) and the first inequality in ( [ e : rur ] ) that @xmath135 for all @xmath134 .
note that @xmath136 as @xmath40 . to get a lower estimate for @xmath137
, we observe that @xmath138 } \nonumber \\ & \quad\exp\left\{-\frac{\varphi(k)^2}{4(\log k+\varphi(k))}\right\ } \nonumber \\ & \le\frac{1}{4\sqrt{\pi}}\frac{(\log k+\varphi(k))^{\frac{3}{2}}}{(\log k)^3 } = \frac{1}{4\sqrt{\pi}}\frac{\left(1+\frac{\phi(k)}{\sqrt{\log k}}\right)^{\frac{3}{2}}}{(\log k)^{\frac{3}{2 } } } \nonumber \\ & \le\frac{1}{4\sqrt{\pi}}\left[\frac{1}{(\log k)^{\frac{3}{2}}}+\frac{3\phi(k)}{2(\log k)^2}+\frac{3\phi(k)^2 } { 8(\log k)^{\frac{5}{2}}}\right],\quad k > k_1 .
\label{e : coc}\end{aligned}\ ] ] in the proof of ( [ e : coc ] ) , we used the inequality @xmath139 it follows from ( [ e : sestim3 ] ) , ( [ e : phi ] ) , ( [ e : rur ] ) , and ( [ e : coc ] ) that @xmath140,\end{aligned}\ ] ] for all @xmath141
. therefore , condition ( [ e : constant ] ) implies that for every @xmath63 there exists a sufficiently large number @xmath111 such that for every @xmath142 , @xmath143 next , taking into account ( [ e : ro1 ] ) , ( [ e : cons ] ) , and the previous inequality , we see that the following statement holds .
[ l : estime ] for every @xmath63 and all @xmath142 , @xmath144 where @xmath145 lemma [ l : estime ] provides estimates for the function @xmath146 , which differ by a quantity of the higher order of smallness than @xmath147 .
let us next choose the function @xmath148 so that the function @xmath149 , given by the formula in ( [ e : fina ] ) with @xmath148 instead of @xmath68 , satisfies @xmath150 .
it follows from ( [ e : fina ] ) that @xmath151 now , using ( [ e : fina4 ] ) we see that in order to find the value of @xmath152 we should solve the following quadratic equation : @xmath153 ^ 2\phi_1(k ) \nonumber \\ & \quad-2\sqrt{\log k}\left[(u_k)^{-1}(g(k))\right]^2=0 .
\label{e : fina5}\end{aligned}\ ] ] solving ( [ e : fina5 ] ) and taking into account ( [ e : fina4 ] ) , we obtain @xmath154 where @xmath114 is defined by ( [ e : fina7 ] )
. we will next check that the function @xmath148 defined by ( [ e : fina6 ] ) is admissible , that is , condition ( [ e : constant ] ) holds for @xmath148 . recall that @xmath150 , and hence ( [ e : r2 ] ) gives @xmath155
it follows from the definition of the function @xmath149 ( see ( [ e : fina ] ) ) that @xmath156 thus @xmath157 it is not hard to see that @xmath148 is a bounded function .
indeed , ( [ e : ogran1 ] ) implies that @xmath158 where @xmath159 is a positive constant .
therefore , @xmath160 and thus the function @xmath148 is bounded for large values of @xmath6 .
now , using ( [ e : ogran1 ] ) , we see that @xmath161 and hence @xmath162 , where @xmath163 is the constant appearing in ( [ e : constant ] ) for the function @xmath148 .
it follows from lemma [ l : estime ] and the equality @xmath150 that @xmath164 where @xmath165 and @xmath166 is given by ( [ e : fina7 ] ) .
therefore , @xmath167 to get a lower estimate for @xmath41 , we will reason similarly .
the only difference here is that we replace the equation @xmath150 by the equation @xmath168 , where @xmath169 is defined by ( [ e : gg ] ) , @xmath63 is fixed , and @xmath170 is a function such as in ( [ e : fina ] ) , but with an unknown function @xmath171 instead of the function @xmath68 .
next , we can prove that @xmath172 where the function @xmath115 is given by ( [ e : final2 ] ) . moreover , the function @xmath171 is bounded for large values of @xmath6 , and @xmath173 therefore @xmath174 , where @xmath175 is the constant appearing in ( [ e : constant ] ) for the function @xmath171 .
let us set @xmath176 with @xmath177 given by ( [ e : final2 ] ) .
it follows that @xmath178 and hence @xmath179 now , it is clear that formula ( [ e : wildi3 ] ) in theorem [ t : ochen ] follows from ( [ e : r1 ] ) , ( [ e : wild1 ] ) , ( [ e : wild2 ] ) , ( [ e : wildi1 ] ) , and ( [ e : wildi2 ] ) .
this completes the proof of theorem [ t : ochen ] .
we will next estimate the difference between the lower and the upper estimates for the implied volatility @xmath8 in formula ( [ e : wildi3 ] ) .
first , note that since the function @xmath148 is eventually bounded , formula ( [ e : fina4 ] ) shows that the function @xmath180 is also eventually bounded .
similarly , the function @xmath181 is eventually bounded .
it follows from ( [ e : fina7 ] ) and ( [ e : final2 ] ) that the functions @xmath182 and @xmath183 , are equivalent to the function @xmath184 near infinity .
therefore , @xmath185 as @xmath40 .
to estimate the difference @xmath186 , we observe that the function @xmath73 is lipschitz on every proper subinterval of the interval @xmath187 with the lipschitz constant independent of @xmath6 on every such interval . now
, it is not hard to see , using ( [ e : gg ] ) , ( [ e : fina7 ] ) , and ( [ e : final2 ] ) that @xmath188\right)=o\left(\frac{1}{\log k}\right)\ ] ] as @xmath40 .
next , by taking into account ( [ e : smale ] ) , we obtain @xmath189 as @xmath40 .
[ t : imvol ] the following formula holds for the implied volatility @xmath24 as @xmath60 : @xmath190 in ( [ e : imvol1 ] ) , the function @xmath114 is defined by ( [ e : fina7 ] ) .
theorem [ t : imvol ] follows from theorem [ t : ochen ] and ( [ e : imvol ] ) .
[ r:1 ] note that the @xmath191-estimate in formula ( [ e : imvol1 ] ) depends on @xmath192 .
more precisely , formula ( [ e : imvol1 ] ) should be understood as follows . for every @xmath63 there
exist @xmath193 and @xmath111 such that @xmath194 for all @xmath195 .
_ proof of theorem [ t : corrf ] ( continuation ) .
expanding the function @xmath196 near zero , we obtain @xmath197 as @xmath198 .
now , using ( [ e : expan ] ) and the fact that the function @xmath199 grows like @xmath184 , we see that @xmath200 as @xmath40 .
moreover , ( [ e : fina7 ] ) and ( [ e : expan ] ) imply that @xmath201 ^ 2 \nonumber \\ & \quad+\sqrt{2}{\sqrt{\log k}}(u_k)^{-1}(g(k))\sqrt{1+\frac{\left[(u_k)^{-1}(g(k))\right]^2 } { 2\log k } } \nonumber \\ & = \sqrt{2}(u_k)^{-1}(g(k))\sqrt{\log k}+\left[(u_k)^{-1}(g(k))\right]^2 \nonumber \\ & \quad+\frac{\sqrt{2}\left[(u_k)^{-1}(g(k))\right]^3}{4\sqrt{\log k } } + o\left((\log k)^{-\frac{3}{2}}\right ) \label{e : imvol3}\end{aligned}\ ] ] as @xmath40 .
next , combining ( [ e : ex1 ] ) and ( [ e : imvol3 ] ) , we see that @xmath202 ^ 2}{2(\log k)^{\frac{1}{2 } } } \nonumber \\ & + \frac{\sqrt{2}\left[(u_k)^{-1}(g(k))\right]^3}{8\log k}-\frac{\left[(u_k)^{-1}(g(k))\right]^2 } { 4(\log k)^{\frac{1}{2}}}-\frac{\sqrt{2}\left[(u_k)^{-1}(g(k))\right]^3}{4\log k } \nonumber \\ & + \frac{\sqrt{2}\left[(u_k)^{-1}(g(k))\right]^3}{8\log k}+o\left((\log k)^{-\frac{3}{2}}\right ) \nonumber \\ & = ( \log k)^{\frac{1}{2}}+\frac{\sqrt{2}}{2}(u_k)^{-1}(g(k))+\frac{\left[(u_k)^{-1}(g(k))\right]^2}{4}(\log k)^{-\frac{1}{2 } } \nonumber \\ & + o\left((\log k)^{-\frac{3}{2}}\right ) \label{e : imvol4}\end{aligned}\ ] ] as @xmath40 .
now , it is clear that ( [ e : imvol2 ] ) follows from ( [ e : imvol1 ] ) and ( [ e : imvol4 ] ) . _
this completes the proof of theorem [ t : corrf ] .
in the present section , we explain how to derive the asymptotic formula for the left wing of the implied volatility due to de marco , hillairet , and jacquier from our formula ( [ e : imvol2 ] ) .
note that formula ( [ e : imvol2 ] ) is very sensitive to even small changes .
such changes often produce errors of order @xmath203 as @xmath60 .
the next statement is essentially the result obtained in theorem 3.7 in @xcite .
[ c : fina ] let @xmath98 .
then @xmath204 in ( [ e : dmhjf ] ) , the function @xmath68 satisfies the following condition : @xmath205 where @xmath206 _ proof .
our first goal is to replace the expression @xmath207 in formula ( [ e : imvol2 ] ) by the expression @xmath208 , and estimate the error . for the sake of shortness
, we put @xmath209 _ [ l : ttu ] let @xmath98 .
then the following asymptotic formula is valid as @xmath60 : @xmath210 where @xmath211}{4\sqrt{t}}(\log u)^{-\frac{1}{2}},\ ] ] and @xmath212 is defined by ( [ e : ppf ] ) .
lemma [ l : ttu ] follows from theorem [ t : corrf ] and ( [ e : ppf ] ) .
the next lemma provides an estimate for the function @xmath212 .
[ l : smallch ] the following formula holds : @xmath213 where the function @xmath214 is given by ( [ e : limsup ] ) .
let us first assume that @xmath215 .
this assumption is equivalent to the following : @xmath87 .
then , using ( [ e : fina3 ] ) , we see that for @xmath216 , we have @xmath217 therefore , for @xmath218 , @xmath219 and @xmath220 since @xmath221 the mean value theorem implies that @xmath222 where @xmath223 next , using ( [ e : r2 ] ) and ( [ e : m ] ) , we see that for every @xmath224 there exists @xmath111 such that @xmath225 moreover , ( [ e : r2 ] ) , ( [ e : m ] ) , and the mean value theorem imply that @xmath226 where the function @xmath227 is positive and satisfies the following condition : @xmath228 _ next , taking into account formulas ( [ e : l1 ] ) - ( [ e : pkk ] ) , we see that lemma [ l : smallch ] holds under the condition @xmath87 .
it remains to prove lemma [ l : smallch ] in the case where @xmath88 .
the previous condition means that @xmath229 .
fix @xmath230 such that @xmath231 .
in addition , fix @xmath6 so large that the following inequalities hold : @xmath232 @xmath233 and @xmath234 next , taking into account ( [ e : i1 ] ) , we assume that @xmath235 then , using ( [ e : i2 ] ) , we obtain @xmath236 in addition , @xmath237 provided that @xmath238 it follows from ( [ e : i1 ] ) and ( [ e : inver1 ] ) that the number @xmath239 satisfies the previous condition .
therefore , @xmath240 moreover , ( [ e : itfo ] ) and the mean value theorem imply that for @xmath241 , @xmath242 where @xmath243 is determined from ( [ e : l1 ] ) , and @xmath244 it is easy to see that @xmath245 therefore , for every @xmath224 there exists @xmath111 such that the inequality in ( [ e : ll ] ) holds for all @xmath142 .
now , the proof of lemma [ l : smallch ] in the case where @xmath88 can be completed exactly as in the case when @xmath87 . finally , it is not hard to see that corollary [ c : fina ] follows from lemmas [ l : ttu ] and [ l : smallch ] .
let us assume @xmath26 and @xmath246 .
we will next compare the numbers @xmath247 and @xmath208 appearing in formula ( [ e : imvol22 ] ) in corollary [ c : sims ] and in the de marco - hillairet - jacquier formula ( [ e : dmhjf ] ) , respectively . recall
that if @xmath87 , then the number @xmath248 is defined for all @xmath246 . on the other hand , if @xmath88 , then @xmath248 is defined under the additional restriction @xmath249 it is clear that if @xmath250 , then @xmath248 and @xmath208 are positive numbers for all @xmath246 . if @xmath251 , then we have @xmath252 and @xmath253 . the remaining case where @xmath254 is interesting . in this case , the number @xmath208 is negative , while the sign of the number @xmath248 can be positive or negative .
we will next clarify the previous statement .
[ l : moreint ] suppose condition ( [ e : spec1 ] ) holds
. then the following are true : 1 .
let the number @xmath25 be such that @xmath255 then @xmath252 .
2 . let the number @xmath25 be such that @xmath256 then @xmath257 .
3 . let the number @xmath25 be such that @xmath258 then @xmath259 .
[ r : price ] note that in the case described in part ( 1 ) of lemma [ l : moreint ] , the numbers @xmath248 and @xmath208 have opposite signs .
the proof of lemma [ l : moreint ] is simple , and we leave it as an exercise for the reader . the next assertion characterizes the limiting behavior of the difference @xmath260 . [ t : estims ] let @xmath261 . then @xmath262=1 .
\label{e : forest}\ ] ] [ r : u ] for @xmath88 , the expression @xmath248 exists if @xmath263 ( see formula ( [ e : accord ] ) ) .
note that for every fixed @xmath90 with @xmath88 , condition ( [ e : suf ] ) holds for sufficiently small values of @xmath6 .
this explains how we should understand formula ( [ e : forest ] ) for @xmath88 .
_ proof of theorem [ t : estims ] .
suppose @xmath87 , and set @xmath264 and @xmath265 _ [ l : aux ] for all @xmath246 , @xmath266 where @xmath267 , @xmath268 , and @xmath269 are given by ( [ e : notka1 ] ) and ( [ e : notka2 ] ) .
_ proof of lemma [ l : aux ] . using the mean value theorem
, we see that @xmath270 where @xmath271 .
it follows from ( [ e : mvt1 ] ) that @xmath272 now , the estimates in lemma [ l : aux ] follow from ( [ e : itfoo ] ) and ( [ e : thef1 ] ) .
_ let us continue the proof of theorem [ t : estims ] .
lemma [ l : aux ] implies that @xmath273 \nonumber \\ & \le\exp\left\{\frac{b_k^2-a^2}{2}\right\}. \label{e : thef2}\end{aligned}\ ] ] since @xmath274 as @xmath60 , formula ( [ e : forest ] ) follows from ( [ e : thef2 ] ) .
the remaining part of the proof of theorem [ t : estims ] resembles that of the second part of lemma [ l : smallch ] .
let us assume @xmath88 .
then we have @xmath229 .
fix @xmath230 such that @xmath275 and suppose @xmath25 is such that @xmath276 and @xmath277 let @xmath278 then we have @xmath279 therefore , @xmath280 provided that @xmath281 since the previous estimates hold for the number @xmath282 , we have @xmath283 . here
@xmath267 and @xmath268 are defined by ( [ e : notka1 ] ) , and @xmath284 next , using ( [ e : ess0 ] ) and ( [ e : ess2 ] ) , we obtain @xmath285 .
it follows from ( [ e : thef1 ] ) and from the inequalities @xmath286 that @xmath287 finally , it is not hard to see that formula ( [ e : forest ] ) with @xmath88 can be derived from the previous estimates and from the equality @xmath288 .
[ r : eshchio ] theorem [ t : estims ] explains why the error term in the de marco - hillairet - jacquier formula is worse than that in formula ( [ e : imvol22 ] ) .
the constant elasticity of variance model ( the cev model ) is described by the following stochastic differential equation : @xmath289 where @xmath290 , @xmath291 , and @xmath292 .
if @xmath293 , then the boundary at @xmath294 is naturally absorbing , while for @xmath295 , we impose an absorbing boundary condition .
the cev model was introduced by j. c. cox in @xcite ( see also @xcite ) .
a useful information about the cev model , including some of the results formulated below , can be found in @xcite .
the cev process is used in the financial industry to model spot prices of equitites and commodities ( see , e.g. , @xcite , and the references therein ) .
fix @xmath20 .
then we have @xmath296 moreover , the density of the absolutely continuous part of the distribution @xmath49 of @xmath297 is as follows : @xmath298 in ( [ e : cev1 ] ) , @xmath299 is the normalized incomplete gamma function given by @xmath300 while in ( [ e : cev2 ] ) , the parameter @xmath301 is defined by @xmath302 , the function @xmath303 is the modified bessel function of the first kind , and the constant @xmath304 is given by @xmath305 it is known that as @xmath198 , @xmath306 for all @xmath307 . therefore , it follows from ( [ e : cev2 ] ) that as @xmath198 , @xmath308 where @xmath309^{\frac{1}{2(1-\rho)}}\gamma\left(\frac{3 - 2\rho}{2(1-\rho)}\right ) } \nonumber \\ & \quad\exp\left\{-\frac{1}{2t\sigma^2(1-\rho)^2}\right\}. \label{e : cev4}\end{aligned}\ ] ] now , taking into account ( [ e : cev3 ] ) and ( [ e : cev4 ] ) , we see that as @xmath60 , @xmath310 therefore , condition ( [ e : ost ] ) holds . finally , applying corollary [ c : sims ] , we derive the following statement .
[ c : cocor ] formula ( [ e : imvol22 ] ) with @xmath90 given by ( [ e : cev2 ] ) holds for the implied volatility in the cev model .
[ r : finn ] propositions , similar to corollary [ c : cocor ] , can be established for many other models besides the cev model , e.g. , jump - to - default models , and models described by processes stopped at the first hitting time of zero . to apply corollary [ c : sims ] to a model with atoms , we only need to know the value of @xmath90 and to estimate the rate of decay of the asset price density near zero .
such information is provided in @xcite for some of the models mentioned above .
the figures included in the present section illustrate the performance of two asymptotic formulas , providing approximations to the left wing of the implied volatility in the cev model : the de marco - hillairet - jacquier formula and formula ( [ e : imvol22 ] ) established in the present paper .
the values of the cev parameters in figures 1 and 2 are chosen as follows : @xmath311 , @xmath312 , @xmath313 , and @xmath314 . under the previous assumptions ,
the value @xmath90 of the mass at zero is approximately equal to @xmath315 . in figures 1 and 2
, the independent variable is the log - moneyness @xmath317 given by @xmath318 .
the large blue stars in figure 1 show the monte carlo estimate of the function @xmath319 to plot the graph of the function represented by blue stars , monte carlo simulations with @xmath320 paths were used , each drawn with 100 time steps .
the solid black curve in figure 1 depicts the full smile approximation using all the three terms in formula ( [ e : imvol22 ] ) .
furthermore , the graph in black dashes corresponds to the smile approximation based on formula ( [ e : imvol22 ] ) with 2 terms , while the graph in black crosses represents the de marco - hillairet - jacquier approximation .
figure 2 shows the approximation errors .
even superficial observations of the graphs in figures 1 and 2 show that formula ( [ e : imvol22 ] ) provides a better approximation to the left wing of the implied volatility in the cev model than the de marco - hillairet - jacquier formula .
note that the graph of the monte carlo estimate of the function defined in ( [ e : mont ] ) and the graph of the approximation to this function based on formula ( [ e : imvol22 ] ) match rather well .
applied mathematics series 55 , national bureau of standards , washington , 1972 .
finance 19 ( 2009 ) , 1 - 12 .
prob . 45 ( 2008 ) , 16 - 32 . in : cont , r. ( ed . ) , _ frontiers in quantitative finance : volatility and credit risk modeling , _ wiley , hoboken , 2009 , 19 - 45 .
preprint , 2010 .
working paper , stanford university , 1975 , ( reprinted in journ .
portfolio management 22 ( 1996 ) , 15 - 17 ) .
financial economics 3 ( 1976 ) , 145 - 166 .
preprint , 2013 , available at arxiv:1310.1020v1 .
futures markets 31 ( 2011 ) , 230 - 250 . to appear in finance and stochastics
; available at ssrn.com/abstract=1768383 .
alternative investments 11 ( 2009 ) , 65 - 84 .
springer verlag berlin heidelberg , 2012 .
finance 15 ( 2012 ) , 1250020 .
siam journ .
math . 1 ( 2010 ) , 609 - 641 .
finance 14 ( 2004 ) , 469 - 480 . | we consider the asymptotic behavior of the implied volatility in stochastic asset price models with atoms . in such models ,
the asset price distribution has a singular component at zero .
examples of models with atoms include the constant elasticity of variance model , jump - to - default models , and stochastic models described by processes stopped at the first hitting time of zero . for models with atoms ,
the behavior of the implied volatility at large strikes is similar to that in models without atoms . on the other hand ,
the behavior of the implied volatility at small strikes is influenced significantly by the atom at zero .
s. de marco , c. hillairet , and a. jacquier found an asymptotic formula for the implied volatility at small strikes with two terms and also provided an incomplete description of the third term . in the present paper
, we obtain a new asymptotic formula for the left wing of the implied volatility , which is qualitatively different from the de marco - hillairet - jacquier formula .
the new formula contains three explicit terms and an error estimate .
we show how to derive the de marco - hillairet - jacquier formula from our formula , and compare the performance of the two formulas in the case of the cev model .
the resulting graphs show that the new formula provides a notably better approximation to the smile in the cev model than the de marco - hillairet - jacquier formula . |
continuous - time jump markov processes are broadly used in stochastic models of operations research . in many applications
continuous - time jump markov processes are defined by transition rates often called @xmath0-functions .
each @xmath0-function defines kolmogorov s backward and forward equations , and transition probabilities of the jump markov process defined by the @xmath0-function satisfy these equations .
if transition rates are unbounded , kolmogorov s equations may have multiple solutions , see , e.g. , anderson ( * ? ? ?
* chap . 4 , example 1.2 ) ,
doob ( * ? ? ?
6 ) , kendall @xcite , reuter @xcite , and the relation between kolmogorov s equations and the corresponding transition probabilities is not trivial . for example , in queueing theory birth and death processes have unbounded transition rates in each of the following three situations : arrival rates depend on the state of the system and are unbounded , queues with an infinite number of servers , queues with reneging
. this paper answers the questions on how a nonhomogeneous jump markov process can be defined for a given @xmath0-function and how can its transition probability be found as a solution of kolmogorov s backward and forward equations .
these questions were studied by feller @xcite for continuous @xmath0-functions and a standard borel state space , by ye et al .
@xcite for measurable @xmath0-functions and a countable state space , and by feinberg et al .
@xcite for measurable @xmath0-functions and a standard borel state space .
all these papers considered @xmath0-functions satisfying certain boundedness conditions .
this paper generalizes the results from feinberg et al .
@xcite to more general classes of unbounded @xmath0-functions , strengthens some of results from @xcite , and provides proofs of the following two facts : ( i ) ( lemma [ l : a - eq](a ) ) fellers s assumption on the boundedness of a @xmath0-function , assumption [ feller ] , is equivalent to the boundedness of a @xmath0-function at each state , assumption [ lb ] , and ( ii ) ( theorem [ thm : intfke ] ) kolmogorov s forward equation is equivalent to the integral equation .
the first fact is introduced and the validity of equation is stated in @xcite without detailed proofs . for a topological space @xmath1 its borel @xmath2-field ( the @xmath2-field generated by open subsets of @xmath3 ) is always denoted by @xmath4 and the sets in @xmath5 are called _ borel subsets _ of @xmath3 .
let @xmath6 be the real line endowed with the euclidean metric . a topological space @xmath7
is called a _ standard borel space _ if there exists a bijection @xmath8 from @xmath7 to a borel subset of @xmath6 such that the mappings @xmath8 and @xmath9 are measurable . in this paper , measurability and borel measurability are used synonymously .
let @xmath10 be a standard borel space , called the state space , and let @xmath11 be a finite or an infinite interval in @xmath12 . in this paper
, we always assume that @xmath13 a function @xmath14 , where @xmath15 , @xmath16u , t_1[$ ] , @xmath17 and @xmath18 is called a transition function if it takes values in @xmath19 $ ] and satisfies the following properties : * for all @xmath20 the function @xmath21 is a measure on @xmath22 ; * for all @xmath23 the function @xmath24 is borel measurable in @xmath25 * @xmath26 satisfies the chapman - kolmogorov equation @xmath27 a transition function @xmath28 is called _ regular _ if @xmath29 for all @xmath20 in the domain of @xmath28 . a stochastic process @xmath30 with values in @xmath31 , defined on the probability space @xmath32 and adapted to the filtration @xmath33
, is called markov if @xmath34 , @xmath35 for all @xmath36 @xmath16u , t_1[$ ] , and @xmath37 .
each markov process has a transition function @xmath28 such that @xmath38 @xmath35 ; see kuznetsov @xcite , where the equivalence of two definitions of a markov process given by kolmogorov @xcite is established .
in addition , if a markov process is a jump process , that is , if each sample path of the process is a right - continuous piecewise constant function in @xmath39 that has a finite or countable number of discontinuity points on @xmath40 , then the markov process is called a _ jump markov process_. a function @xmath41 , where @xmath42 , @xmath40 , and @xmath43 , is called a _ q - function _ if it satisfies the following properties : * for all @xmath44 the function @xmath45 is a signed measure on @xmath22 such that @xmath46 @xmath47 @xmath48 and @xmath49 for all @xmath50 ; * for all @xmath23 the function @xmath41 is measurable in @xmath51 in addition to properties ( a ) and ( b ) , if @xmath52 for all @xmath44 , then the @xmath0-function @xmath53 is called _
conservative_. note that any @xmath0-function can be transformed into a conservative @xmath0-function by adding an absorbing state @xmath54 to @xmath31 with @xmath55 , @xmath56 , and @xmath57 , where @xmath42 and @xmath58 . to simplify the presentation , in this paper
we always assume that @xmath53 is conservative . the same arguments as in remark 4.1 in feinberg et al .
@xcite explain how the main formulations change when the @xmath0-function @xmath53 is not conservative .
a @xmath0-function @xmath53 is called _ continuous _ if it is continuous in @xmath40 .
feller @xcite studied kolmogorov s backward and forward equations for continuous @xmath0-functions and provided explicit formulae for a transition function that satisfies kolmogorov s backward and forward equations . if the constructed transition function is regular , feller ( * ? ? ?
* theorem 3 ) showed that this transition function is the unique solution of kolmogorov s backward equation .
though feller @xcite focused on regular transition functions , it follows from the proof of theorem 3 in feller @xcite that the transition function constructed there is the minimal solution of kolmogorov s backward equation .
feinberg et al .
@xcite showed for a measurable @xmath0-function that the transition function constructed by feller @xcite is the minimal solution of kolmogorov s backward and forward equations , and it is the transition function of the jump markov process defined by the random measure whose compensator is defined via the @xmath0-function . in this paper , we show that the minimal solution of kolmogorov s backward and forward equations is the transition function of the corresponding jump markov process under more general boundedness assumptions on @xmath0-functions than those assumed in @xcite .
in this section , we describe several assumptions on unbounded @xmath0-functions and the results of this paper .
let @xmath59 for @xmath42 and @xmath60 and let @xmath61 for @xmath62 feller @xcite studied kolmogorov s equations for continuous @xmath0-functions under the following assumption .
[ feller ] there exists borel subsets @xmath63 of @xmath31 such that @xmath64 for all @xmath65 and @xmath66 as @xmath67 .
feinberg et al .
@xcite studied kolmogorov s equations for measurable @xmath0-functions under the following assumption with @xmath68 and @xmath69 .
[ lb ] @xmath70 for each @xmath42 . as mentioned in feinberg et al .
@xcite , assumptions [ feller ] and [ lb ] are equivalent ; see lemma [ l : a - eq](a ) for details . in this section ,
we introduce two more general assumptions .
[ alb ] @xmath71 for each @xmath72t_0,t_1[$ ] and @xmath62 [ l1 ] @xmath73 for each @xmath74t_0 , t_1[$ ] and @xmath75 the following lemma compares assumptions [ feller][l1 ] .
[ l : a - eq ] the following statements hold for a measurable @xmath0-function @xmath76 \(a ) assumptions [ feller ] and [ lb ] are equivalent ; \(b ) assumption [ lb ] implies assumption [ alb ] ; \(c ) assumption [ alb ] implies assumption [ l1 ] .
\(a ) let @xmath77 be a sequence of borel subsets of @xmath31 satisfying the properties stated in assumption [ feller ] .
then for each @xmath42 there exists an @xmath78 such that @xmath79 and therefore @xmath80 .
thus , assumption [ feller ] implies assumption [ lb ] . to prove that assumption [ lb ] implies assumption [ feller ] , define @xmath81 since @xmath82 are countable intersections of projections of borel sets , the sets @xmath83 are analytic for all @xmath84 see bertsekas and shreve ( * ? ? ? * proposition 7.39 and corollary 7.35.2 ) .
in addition , assumption [ lb ] implies that @xmath85 thus , in view of the novikov separation theorem , kechris ( * ? ? ? * theorem 28.5 ) , there exist borel subsets @xmath86 @xmath87 of @xmath31 such that @xmath88 and @xmath89 this fact implies that @xmath90 and @xmath91 where the sets @xmath92 and @xmath93 are complements of the sets @xmath94 and @xmath83 , respectively .
let @xmath95 for all @xmath96 the borel sets @xmath97 @xmath87 satisfy the properties stated in assumption [ feller ] .
( b , c ) statements ( b ) and ( c ) are obvious .
[ t - eq ] under assumption [ feller ] , which , as stated in lemma [ l : a - eq](a ) , is equivalent to assumption [ lb ] , feller @xcite studied kolmogorov s equations for the time parameter @xmath98 under assumption [ lb ] , feinberg et al .
@xcite studied kolmogorov s equations for the time parameter @xmath99 it is apparent that the formulation of results for an arbitrary interval @xmath100 where @xmath101 is more general than their formulation for the interval @xmath102 in fact , these two formulations are equivalent under assumption [ lb ] holding for the corresponding time intervals .
indeed , a @xmath0-function @xmath103 defined for @xmath104 and satisfying assumption [ lb ] on this interval , can be extended to all @xmath105 by setting @xmath106 for @xmath107 @xmath108 and @xmath109 the extended @xmath0-function satisfies assumption [ lb ] for @xmath110 since solutions of kolmogorov s equations and for the extended @xmath0-function are constants in @xmath111 when @xmath112 and @xmath113 and since kolmogorov s equations for the original @xmath0-function @xmath53 and the extended @xmath0-function coincide when @xmath60 there is a one - to - one correspondence between solutions of kolmogorov s equations for the @xmath0-function @xmath53 and for the extended @xmath0-function . since assumption [ lb ] is assumed in @xcite , the results obtained in @xcite for the problem formulations for the interval @xmath114 hold for an arbitrary interval @xmath115 as explained in remark [ t - eq ] , if a @xmath0-function @xmath53 satisfies assumption [ lb ] on the interval @xmath100 its extension to @xmath105 defined in remark [ t - eq ] satisfies the same assumption on @xmath116 the following example illustrates that this need not be the case if @xmath53 satisfies assumption [ alb ] or assumption [ l1 ] .
hence , following feller @xcite , we formulate the results in this paper for an arbitrary interval @xmath11 with @xmath117 [ ex : b - eq ] _ a @xmath0-function @xmath53 satisfies assumption [ alb ] on the interval @xmath100 while its extention to @xmath105 defined in remark [ t - eq ] does not satisfy even the weaker assumption [ l1 ] when @xmath68 and @xmath69 .
_ fix an arbitrary @xmath1180 , \infty[.$ ] let @xmath119 , @xmath120 and @xmath121 for all @xmath122 . then @xmath123 for each @xmath72t_0,t_1[$ ] and @xmath62
thus the @xmath0-function @xmath53 satisfies assumption [ alb ] .
consider the extension of @xmath53 to @xmath105 defined in remark [ t - eq ] and the sequence @xmath124 with @xmath125 for all @xmath126 observe that @xmath127 therefore , the described extension of @xmath53 from @xmath128 to @xmath105 does not satisfy assumption [ l1 ] . in section
[ s - be ] we show in theorem [ thm : jmp ] that under assumption [ l1 ] the compensator defined by a @xmath0-function and an initial probability measure define a jump markov process , whose transition function @xmath129 is described in , and theorem [ thm : bke ] states that this function is the minimal function satisfying kolmogorov s backward equation .
the function @xmath129 was introduced in feller @xcite .
section [ s - fe ] deals with kolmogorov s forward equation , when assumption [ alb ] holds , and theorem [ thm : fkde ] states that @xmath129 is the minimal function satisfying the forward equation .
section [ s - gb ] presents results on kolmogorov s forward equation under assumption [ lb ] .
in this section , we show that a @xmath0-function satisfying assumption [ l1 ] defines a transition function for a jump markov process .
in addition , this transition function is the minimal function satisfying kolmogorov s backward equation defined by this @xmath0-function .
let @xmath130 be an isolated point adjoined to the space @xmath31 .
denote @xmath131.consider the borel @xmath2-field @xmath132 on @xmath133 , which is the minimal @xmath2-field containing @xmath134 and @xmath135 let @xmath136t_0 , t_1])^\infty$ ] be the set of all sequences @xmath137 with @xmath138 and @xmath139t_0 , t_1]$ ] for all @xmath140 this set is endowed with the @xmath2-field generated by the products of the borel @xmath2-fields @xmath141 and @xmath142t_0 , t_1])$ ]
. denote by @xmath143 the subset of all sequences @xmath144 from @xmath136t_0 , t_1])^\infty$ ] such that : ( i ) @xmath145 ; ( ii ) for all @xmath146 if @xmath147 , then @xmath148 and @xmath149 , and if @xmath150 , then @xmath151 and @xmath152 .
observe that @xmath143 is a measurable subset of @xmath136t_0 , t_1])^\infty$ ] .
consider the measurable space @xmath153 , where @xmath154 is the @xmath2-field of the measurable subsets of @xmath143 .
for all @xmath155 , let @xmath156 and @xmath157 where @xmath158 be the random variables defined on the measurable space @xmath159 .
let @xmath160 , @xmath161 , @xmath162 , and for all @xmath163,$ ] let @xmath164 , where @xmath165 throughout this paper , we omit @xmath166 whenever possible .
consider the multivariate point process @xmath167 on @xmath159 . given
a @xmath0-function @xmath53 satisfying assumption [ l1 ] , define a random measure @xmath168 on @xmath169 as @xmath170 , b ) : = \int_{t_0}^{t}\sum_{n \ge 0}i\{t_n < s \leq t_{n+1}\ } q(x_n , s , b \setminus \{x_n\})ds , \quad t \in [ t_0 , t_1[,\ b \in { \mathfrak{b}}({\bf x}).\ ] ] observe that @xmath168 is a predictable random measure . indeed , formula ( [ compensator ] ) coincides with feinberg et al .
* eq . ( 2 ) ) when @xmath68 and @xmath69 .
arguments similar to those following feinberg et al .
* eq . ( 2 ) ) , which show that the random measure @xmath168 defined in ( * ? ?
( 2 ) ) is a predictable random measure , imply that the measure @xmath168 defined in is a predictable random measure .
furthermore , @xmath171 for all @xmath172t_0,t_1[$ ] and @xmath173 according to jacod ( * ? ? ?
* theorem 3.6 ) , the predictable random measure @xmath168 defined in and a probability measure @xmath174 on @xmath175 define a unique probability measure @xmath176 on @xmath159 such that @xmath177 and @xmath168 is the compensator of the random measure of the multivariate point process @xmath178 defined by the triplet @xmath32 .
consider the process @xmath179 , @xmath180 defined on @xmath32 and adapted to the filtration @xmath181 . by definition ,
the process @xmath182 is a jump process . for @xmath42 and @xmath60
let @xmath183 be the measure on @xmath184 with values @xmath185 @xmath43 . in this paper
, we use the notation @xmath186 following feller ( * ? ? ?
* theorem 2 ) , for @xmath42 , @xmath15 , @xmath16u , t_1[$ ] , and @xmath43 , define @xmath187 and @xmath188 set @xmath189 according to feller ( * ? ? ? * ( 27 ) and theorem 4 ) , equation can be rewritten as @xmath190 though feller @xcite considered continuous @xmath0-functions , the proof of given in feller ( * ? ? ?
* theorem 4 ) remains correct for measurable @xmath0-functions .
observe that @xmath191 is a transition function if the @xmath0-function @xmath53 satisfies assumption [ l1 ] . for continuous @xmath0-functions satisfying assumption [ feller ] , feller (
* theorems 2 , 5 ) proved that : ( a ) for fixed @xmath20 the function @xmath192 is a measure on @xmath184 such that @xmath193 , and ( b ) for all @xmath194 the function @xmath195 satisfies the chapman - kolmogorov equation .
the proofs remain correct for measurable @xmath0-functions satisfying assumption [ l1 ] .
the measurability of @xmath195 in @xmath20 for all @xmath43 follows directly from the definitions , , and .
therefore , if @xmath53 satisfies assumption [ l1 ] , the function @xmath191 takes values in @xmath19 $ ] and satisfies properties ( i)-(iii ) from the definition of a transition function .
feinberg et al .
* theorem 2.2 ) ) [ thm : jmp ] given a probability measure @xmath174 on @xmath31 and a @xmath0-function @xmath53 satisfying assumption [ l1 ] , the jump process @xmath196 defined in is a jump markov process with the transition function @xmath191 .
the statement of the theorem follows from the same arguments as in the proof of feinberg et al .
* theorem 2.2 ) , where the case @xmath197 and @xmath198 was considered .
we remark that though it was assumed there that the @xmath0-function @xmath53 satisfies assumption [ lb ] , the arguments in the proof in @xcite only require that , @xmath199u , t_1[,\ ] ] and this holds in view of assumption [ l1 ] . let @xmath200 be the family of all real - valued non - negative functions @xmath201 defined for all @xmath16t_0 , t_1[,$ ] @xmath202 @xmath17 and @xmath18 which are measurable in @xmath203 for all @xmath172t_0 , t_1[$ ] and @xmath204 observe that @xmath205 consider a set @xmath206 and some family @xmath207 of functions @xmath208.$ ] a function @xmath8 from @xmath207 is called minimal in the family @xmath207 if for every function @xmath209 from @xmath207 the inequality @xmath210 holds for all @xmath211 the following theorem generalizes theorems 3.1 and 3.2 in feinberg et al .
@xcite , stating the same statements under assumption [ lb ] which is stronger than assumption [ l1 ] .
the measurability in @xmath212 of a function satisfying kolmogorov s backward equation is implicitly assumed in feinberg et al .
@xcite .
[ thm : bke ] under assumption [ l1 ] , the transition function @xmath191 is the minimal function in @xmath200 satisfying the following two properties : \(i ) for all @xmath213t_0,t_1[,$ ] @xmath107 and @xmath214 @xmath215 and the function is absolutely continuous in @xmath216 \(ii ) for all @xmath213t_0,t_1[,$ ] @xmath107 and @xmath214 kolmogorov s backward equation @xmath217 holds for almost every @xmath218 in addition , if the transition function @xmath191 is regular ( that is , @xmath219 @xmath220 for all @xmath221 @xmath222 @xmath39 in the domain of @xmath129 ) , then @xmath191 is the unique function in @xmath200 satisfying properties ( i ) , ( ii ) and which is a measure on @xmath10 for all @xmath213t_0,t_1[,$ ] @xmath223 and @xmath107 and taking values in @xmath19 $ ] . under assumption [ lb
] , this theorem is theorems 3.1 and 3.2 from feinberg et al . @xcite combined .
however , the proofs there only use the property that holds , and this property is true under assumption [ l1 ] .
therefore , the statement of the theorem holds .
under assumption [ lb ] , kolmogorov s forward equation was studied by feller ( * ? ? ? *
theorem 1 ) for continuous @xmath0-functions and by feinberg et al .
* theorems 4.1 , 4.3 ) for measurable @xmath0-functions . in this section ,
we study kolmogorov s forward equation under assumption [ alb ] , which , in view of lemma [ l : a - eq](b ) , is more general than assumption [ lb ] .
let @xmath224 be the family of real - valued functions @xmath225 defined for all @xmath15 , @xmath16u , t_1[$ ] , @xmath17 and @xmath18 which are measures on @xmath10 for fixed @xmath221 @xmath222 @xmath39 and are measurable functions in @xmath39 for fixed @xmath221 @xmath222 @xmath226 in particular , @xmath227 where @xmath129 is defined in .
[ defxnt ] for @xmath74t_0 , t_1],$ ] a set @xmath228 is called @xmath229-bounded if the function @xmath230 is bounded on the set @xmath231 [ defbbf ] a @xmath232-bounded set is called @xmath53-bounded . in definition [ defbbf ]
we follow the terminology from feinberg et al .
feller @xcite called such sets _ bounded_. the following theorem shows that the transition function @xmath191 is the minimal function satisfying kolmogorov s forward equation .
being applied to a function @xmath53 satisfying the stronger assumption [ lb ] , this theorem implies corollary [ corr2 ] , which is a stronger result than ( * ? ? ?
* theorem 4.3 ) ; see explanations before corollary [ corr2 ] .
[ thm : fkde ] under assumption [ alb ] , the transition function @xmath191 is the minimal function in @xmath224 satisfying the following two properties : \(i ) for all @xmath36 @xmath74u , t_1[,$ ] @xmath17 and @xmath229-bounded sets @xmath233 @xmath234 and the function is absolutely continuous in @xmath16u , s[;$ ] \(ii ) for all @xmath36 @xmath74u , t_1[,$ ] @xmath17 and @xmath229-bounded sets
@xmath233 kolmogorov s forward equation @xmath235 holds for almost every @xmath16u , s[.$ ] in addition , if the transition function @xmath191 is regular , then @xmath191 is the unique function in @xmath224 satisfying properties ( i ) , ( ii ) and taking values in @xmath19 $ ] .
as stated in theorem [ thm : fkde ] , the function @xmath191 satisfies kolmogorov s forward equation for @xmath229-bounded sets @xmath43 . in general , as the following example demonstrates , it is not possible to extend to all sets @xmath43 .
[ ex ] _ for a set @xmath214 kolmogorov s forward equation does not hold at all @xmath16u , t_1[$ ] .
_ let @xmath236 where @xmath237 denotes the set of integers , @xmath238 for all @xmath239 and @xmath240 for all @xmath241 if @xmath242 then starting at time @xmath243 the process spends an exponentially distributed amount of time at state 0 , then it jumps to a state @xmath244 with probability @xmath245 and then it oscillates between the states @xmath246 and @xmath247 with equal intensities .
thus for all @xmath15 and @xmath16u , t_1[$ ] @xmath248 which implies @xmath249 thus , if @xmath250 , then does not hold with @xmath251 because both integrals in are infinite .
the following theorem describes the necessary and sufficient condition for a function @xmath28 from @xmath224 to satisfy properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] . in other words
, it provides a necessary and sufficient condition that a function @xmath28 from @xmath224 satisfies kolmogorov s forward equation .
the necessity part of this theorem plays the central role in proving the minimality property of @xmath129 stated in theorem [ thm : fkde ] .
[ thm : intfke ] let assumption [ alb ] hold .
a function @xmath28 from @xmath224 satisfies properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] if and only if , for all @xmath36 @xmath16u , t_1[,$ ] @xmath17 and @xmath18 @xmath252 [ lem : fkde - sol ] under assumption [ alb ] , the following statements hold : \(a ) for each @xmath15 , @xmath72u , t_1[,$ ] @xmath42 , and @xmath43 , the function @xmath253 satisfies the boundary condition and is absolutely continuous in @xmath16u , s[.$ ] \(b ) the function @xmath191 satisfies property ( ii ) stated in theorem [ thm : fkde ] .
\(a ) under assumption [ lb ] , statement ( a ) of this lemma is theorem 4.1(i ) in feinberg et al .
@xcite , and the proof there is correct if holds . in view of lemma
[ l : a - eq](c ) , formula is true under assumption [ alb ] , and therefore , statement ( a ) of the lemma holds .
\(b ) fix an arbitrary @xmath74t_0 , t_1[.$ ] observe that a @xmath0 function satisfying assumption [ alb ] satisfies assumption [ lb ] with @xmath254 .
then it follows from feinberg et al .
* theorem 4.1(ii ) ) that , for all @xmath255 , @xmath42 , and @xmath229-bounded sets @xmath43 , the function @xmath195 satisfies kolmogorov s forward equation for almost every @xmath16u , s[.$ ] since @xmath256 was chosen arbitrarily , this fact implies that the function @xmath191 satisfies property ( ii ) stated in theorem [ thm : fkde ] . to prove theorems
[ thm : fkde ] and [ thm : intfke ] we formulate and prove two lemmas .
lemmas [ cor ] and [ l : int - s ] present kolmogorov s forward equation in integral forms , , which are equivalent to its differential form .
in particular , theorem [ thm : intfke ] follows from lemma [ l : int - s ] .
let @xmath36 @xmath74u , t_1[$ ] , @xmath257 and @xmath43 be a @xmath229-bounded set . for any function @xmath28 from @xmath258 @xmath259u , s[}q(y , t)\right ) p(u , x;t , b ) < \infty , \qquad t \in ] u , s[.\ ] ] in addition , for @xmath221 @xmath260 @xmath222 and @xmath23 described above , if the function @xmath28 satisfies the boundary condition and is absolutely continuous in @xmath16u , s[,$ ] then it is bounded in @xmath16u , s[,$ ] which along with implies that @xmath261u , s[.\ ] ] [ cor ] for arbitrary fixed @xmath15 , @xmath74u , t_1[$ ] , @xmath42 , and @xmath229-bounded set @xmath18 a function @xmath28 from @xmath224 satisfies the equality @xmath262u , s [ , \end{split}\ ] ] if and only if it satisfies the boundary condition , is absolutely continuous in @xmath16u , s[,$ ] and satisfies kolmogorov s forward equation for almost every @xmath172u , s[.$ ] suppose that a function @xmath28 from @xmath224 satisfies the boundary condition , is absolutely continuous in @xmath16u , s[,$ ] and satisfies kolmogorov s forward equation for almost every @xmath172u , s[$ ] for @xmath221 @xmath260 @xmath222 and @xmath23 described in the formulation of the lemma . since every absolutely continuous function is the integral of its derivative , equality follows from integrating equation from @xmath243 to @xmath39 and using the boundary condition . in particular
, both integrals in equality are finite because , in view of , the first integral is finite . now , suppose holds for @xmath221 @xmath260 @xmath222 and @xmath23 described in the formulation of the lemma .
observe that , for fixed @xmath221 @xmath260 @xmath222 and @xmath23 , the real valued function @xmath26 is a constant plus the the difference of two integrals from @xmath243 to @xmath39 of nonnegative integrable functions defined for @xmath263u , s[$ ] .
since an integral of an integrable function is an absolutely continuous function of the upper limit of integration and its derivative is equal to the integrand almost everywhere on its domain ( royden ( * ? ? ?
* thms 10 on p. 107 and 14 on p. 110 ) ) , the function @xmath26 is absolutely continuous in @xmath16u , s[,$ ] and kolmogorov s forward equation holds for almost every @xmath172u , s[$ ] for the fixed @xmath221 @xmath260 @xmath222 and @xmath23 .
in addition , the absolute continuity of the integrals in implies that holds .
[ l : int - s ] let @xmath15 , @xmath74u , t_1[$ ] , @xmath42 , and @xmath264 be a @xmath229-bounded set .
a function @xmath28 from @xmath224 satisfies for all @xmath265 the boundary condition , is absolutely continuous in @xmath16u , s[,$ ] and satisfies kolmogorov s forward equation for almost every @xmath172u , s[$ ] if and only if it satisfies equality for all @xmath16u , s[$ ] and @xmath266 [ t - eq2 ] the sufficiency statements of theorem [ thm : intfke ] and lemma [ l : int - s ] are not used in the proofs in this section .
_ of lemma
[ l : int - s ] _ the following version of fubini s theorem from halmos ( * ? ? ?
* section 36 , remark ( 3 ) ) is used in the proof .
let @xmath267 be a measure space with @xmath268 and let @xmath269 be a measurable space .
suppose that to almost every @xmath270 there corresponds a finite measure @xmath271 on @xmath272 such that the function @xmath273 is measurable in @xmath274 for each measurable subset @xmath23 of @xmath275 then , for any non - negative measurable function @xmath209 on @xmath276 , @xmath277 where , for each measurable subset @xmath23 of @xmath276 , @xmath278 let us fix @xmath279 @xmath15 , @xmath74u , t_1[$ ] , @xmath42 , and a @xmath229-bounded set @xmath264 . to simplify notations , define @xmath280u , s[,\ b \in { \mathfrak{b}}(c),\\ \label{g-2 } g^{(2)}(t , b )
& : = \int_{\bf x } { \delta}_z(b)q(z , t ) { p}(u , x;t , dz),\qquad & t \in ] u , s[,\ b \in { \mathfrak{b}}(c),\end{aligned}\ ] ] where @xmath281 is the dirac measure on @xmath282 @xmath283 observe that , for @xmath284 the function @xmath285 is a measure on @xmath286 for every @xmath16u , s[,$ ] and @xmath287 is a measurable function on @xmath288u , s[$ ] for every @xmath266 let @xmath16u , s[,$ ] @xmath289u , t[,$ ] and @xmath266 consider @xmath290 and @xmath291 for @xmath292 which is finite for all @xmath270 since @xmath53 is a @xmath0-function , and for @xmath293 formula yields @xmath294 _ necessity .
_ for all @xmath265 , let the function @xmath28 satisfy the boundary condition , be absolutely continuous in @xmath16u , s[,$ ] and satisfy kolmogorov s forward equation for almost every @xmath172u , s[.$ ] equation can be rewritten as @xmath295 formula means that @xmath296 for all @xmath172u , s[.$ ] this inequality and imply that , for @xmath284 @xmath297u , s[.\ ] ] for @xmath284 consider the non - negative functions @xmath298u , s [ \times { \mathfrak{b}}(c))\to { \mathbb}{r}_+,$ ] @xmath299u , s[,\ b \in { \mathfrak{b}}(c).\ ] ] in view of lemma [ cor ] , @xmath300u , s[,\ b \in { \mathfrak{b}}(c).\ ] ] equality , which implies for @xmath301 and yield @xmath302u , s[,\ b \in { \mathfrak{b}}(c).\ ] ] observe that , for any measure @xmath303 on @xmath286 and @xmath304 , @xmath305 where the first equality is correct since @xmath306 and the last one is obtained by changing the order of integration in @xmath307 and @xmath308 and applying fubini s theorem .
let @xmath284 @xmath16u , s[$ ] , and @xmath266 then @xmath309 where the first equality follows from , the second equality follows from with @xmath310 the third equality is obtained by changing the order of integration in @xmath311 and @xmath308 , and the last one is obtained from formula by setting @xmath312u , v[,{\mathfrak{b}}(]u , v[),\lambda),$ ] where @xmath313 is the lebesgue measure , @xmath314 @xmath315 which , in view of inequality is finite for almost every @xmath270 , and @xmath316 for @xmath289u , t[$ ] , by setting @xmath317 , @xmath318 , @xmath319 , and @xmath320 formula yields @xmath321 therefore , for all @xmath16u , s[$ ] and @xmath322 @xmath323 where the first equality follows from with @xmath324 , the second equality follows from and , the third equality follows from , , and , the fourth equality follows from with @xmath325 and @xmath326 and with @xmath327 , and the last one follows from .
thus , and imply . _ sufficiency .
_ assume that the function @xmath328 satisfies for all @xmath16u , s[$ ] and @xmath265 .
as follows from lemma [ cor ] , it is sufficient to show that holds for all @xmath266 in view of equality , formula can be rewritten as @xmath329 let @xmath16u , s[$ ] and @xmath266 since @xmath330 it follows from and fubini s theorem that , for any measure @xmath303 on @xmath331 @xmath332 observe that formula differs from .
next , @xmath333 where the first equality follows from with @xmath334 , the second equality is obtained by interchanging the order of integration in @xmath311 and @xmath308 , and the last one is obtained from by setting @xmath335u , v[,{\mathfrak{b}}(]u , v[),\lambda),$ ] where @xmath313 is the lebesgue measure , @xmath318 , @xmath336 , which , in view of equality and the property that the function @xmath28 takes values in @xmath337 is finite for @xmath338 and for almost every @xmath339 , and @xmath340 therefore , @xmath341 where the first equality follows from , with @xmath325 , and , and the last one is obtained by substituting @xmath342 with .
thus , it follows from and the above equality that holds for all @xmath266 _ of theorem [ thm : intfke ] _ the sufficiency statement of the theorem follows immediately from lemma [ l : int - s ] , and the necessity statement of the theorem follows from lemma [ l : int - s ] and lebesgue s monotone convergence theorem , as explained below
. _ necessity .
_ assume that , for all @xmath36 @xmath74u , t_1[,$ ] @xmath42 , and @xmath229-bounded sets @xmath343 , properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] hold for the function @xmath28 .
assumption [ alb ] and lemma [ l : a - eq](a ) imply that for each @xmath72t_0,t_1 [ $ ] there exist @xmath229-bounded sets @xmath344 such that @xmath345 as @xmath67 .
then , for all @xmath36 @xmath74u , t_1[,$ ] @xmath16u , s[,$ ] @xmath42 , and @xmath18 @xmath346 where the first equality is correct since the sets @xmath347 as @xmath67 , the second equality follows from lemma [ l : int - s ] , and the last one follows from lebesgue s monotone convergence theorem since the sets @xmath347 as @xmath67 .
since the above equality holds for all @xmath16u , s[$ ] for each @xmath74u , t_1[$ ] , formula holds for all @xmath36 @xmath16u , t_1[,$ ] @xmath42 , and @xmath348 _ of theorem [ thm : fkde ] _ in view of lemma [ lem : fkde - sol ] , we need to prove only the minimality and uniqueness properties of @xmath129 among functions from @xmath224 satisfying properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] .
let @xmath349 be a function from @xmath224 satisfying these properties .
let @xmath36 @xmath16u , t_1[,$ ] @xmath42 , and @xmath348 in view of theorem [ thm : intfke ] , formula holds . since the last term in is non - negative , @xmath350 where the last equality is .
assume that for some @xmath351 @xmath352 then , from , , and , @xmath353 thus , by induction , holds for all @xmath354 let @xmath355 then and imply that @xmath356 therefore , the function @xmath191 is the minimal function from @xmath224 satisfying properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] . in conclusion , let the transition function @xmath129 be regular .
if there is another function @xmath357 which satisfies properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] and takes values in @xmath19 $ ] , then , since @xmath129 is the minimal solution , @xmath358 for some @xmath36 @xmath17 @xmath16u , t_1[,$ ] and @xmath348 in addition , @xmath359 therefore , @xmath360 and the inequality @xmath361 contradicts the property that @xmath328 takes values in @xmath19.$ ] theorems [ thm : fkde ] and [ thm : intfke ] imply the following two corollaries .
[ cor - g ] under assumption [ alb ] , the following statements hold : \(a ) for all @xmath36 @xmath74u , t_1[,$ ] @xmath17 and @xmath229-bounded sets @xmath233 the function + @xmath195 satisfies .
\(b ) the function @xmath191 is the minimal function in @xmath224 for which statement ( a ) holds .
in addition , if the transition function @xmath191 is regular , then @xmath191 is the unique function in @xmath224 with values in @xmath19 $ ] for which statement ( a ) holds . in view of lemma [ cor ] , any function @xmath28 from @xmath362 satisfies statement ( a ) of the corollary if and only if it satisfies properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] .
thus , the corollary follows from theorem [ thm : fkde ] .
[ cor : intfke ] let assumption [ alb ] hold .
the function @xmath129 is the minimal function @xmath28 in @xmath224 satisfying equality for all @xmath36 @xmath16u , t_1[,$ ] @xmath17 and @xmath348 in addition , if the transition function @xmath191 is regular , then @xmath191 is the unique function in @xmath224 with values in @xmath19 $ ] satisfying equality for all @xmath36 @xmath16u , t_1[,$ ] @xmath17 and @xmath348 the corollary follows from theorems [ thm : fkde ] and [ thm : intfke ] .
this section provides additional results on kolmogorov s forward equation when assumption [ lb ] holds . under assumption [ lb ]
kolmogorow s forward equation is studied in feinberg et al .
* theorems 4.1 , 4.3 ) , and corollary [ corr2 ] is a more general statement than ( * ? ? ?
* theorem 4.3 ) .
in addition , corollary [ cor - p ] describes the minimality property of the function @xmath363 that is useful for applications to continuous - time markov decision processes .
the following lemma and its corollary do not require any of assumptions [ feller]-[l1 ] .
[ lemmabt ] let @xmath364 @xmath365 and @xmath23 be a @xmath53-bounded set .
a function @xmath366 satisfies kolmogorov s forward equation for almost every @xmath172u , s[$ ] for all @xmath72u , t_1[$ ] if and only if it satisfies this equation for almost every @xmath172u , t_1[.$ ] the sufficiency statement of the lemma is straightforward since @xmath288u , s[\subset ] u , t_1[$ ] when @xmath72u , t_1[.$ ] let a function @xmath366 satisfy kolmogorov s forward equation for almost every @xmath172u , s[$ ] for all @xmath72u , t_1[,$ ] where @xmath36 @xmath17 and @xmath43 is a @xmath53-bounded set .
consider an arbitrary sequence @xmath367 as @xmath368 with @xmath369 let @xmath276 be the set of all @xmath172u , t_1[$ ] such that does not hold at point @xmath370 for @xmath371 the lebesgue measure of the sets @xmath372u , s_n[$ ] is 0 since each of these sets consists of points @xmath172u , s_n[$ ] at which does not hold .
this implies that the lebesgue measure of the set @xmath276 is @xmath48 .
therefore , the function @xmath28 satisfies kolmogorov s forward equation for almost every @xmath16u , t_1[.$ ] [ lemmab ] a function @xmath366 satisfies for @xmath53-bounded sets @xmath23 properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] if and only if the following two properties hold : ( a ) for all @xmath15 , @xmath42 , and @xmath53-bounded sets @xmath23 , the function @xmath26 satisfies the boundary condition and is absolutely continuous in @xmath16u , s[$ ] for each @xmath74u , t_1[;$ ] ( b ) for all @xmath15 , @xmath42 , and @xmath53-bounded set @xmath23 , the function @xmath26 satisfies kolmogorov s forward equation for almost every @xmath172u , t_1[.$ ] for @xmath53-bounded sets @xmath233 property ( i ) stated in theorem [ thm : fkde ] coincides with property ( a ) stated in the corollary .
lemma [ lemmabt ] implies that property ( ii ) stated in theorem [ thm : fkde ] holds for a @xmath53-bounded set @xmath23 if and only if property ( b ) stated in the corollary holds .
let the function @xmath28 satisfy properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] .
since a @xmath53-bounded set is @xmath229-bounded , it follows from corollary [ lemmab ] that properties ( a ) and ( b ) stated in corollary [ lemmab ] hold .
let properties ( a ) and ( b ) stated in corollary [ lemmab ] hold .
fix arbitrary @xmath15 , @xmath74u , t_1[$ ] , and @xmath75 lemma [
l : int - s ] implies that for every @xmath53-bounded set @xmath23 equality holds for all @xmath172u , s[.$ ] in view of assumption [ lb ] and lemma [ l : a - eq](a ) , there exist @xmath53-bounded sets @xmath373 such that @xmath374 @xmath87 and @xmath375 let @xmath376 .
then @xmath377 @xmath87 are @xmath53-bounded sets .
therefore , for each set @xmath378 equality holds for all @xmath16u , s[.$ ] since @xmath379 as @xmath380 lebesgue s monotone convergence theorem implies that this formula also holds for @xmath226 thus , in view of theorem [ thm : intfke ] , the function @xmath28 satisfies properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] .
the following corollary generalizes feinberg et al .
* theorem 4.1 ) since assumption [ alb ] is weaker than assumption [ lb ] .
we remark that absolute continuity in @xmath172u,\infty[$ ] in ( * ? ? ? * theorem 4.1(i ) ) is meant in the sense that for each @xmath72u,\infty[$ ] the function is absolutely continuous in @xmath172u , s[.$ ] for @xmath198 this is equivalent to the absolutely continuity assumed in property ( a ) stated in corollary [ lemmab ] . for unbounded intervals , this type of absolute continuity
is sometimes called local absolute continuity .
feinberg et al .
* theorem 4.1 ) ) [ corr1 ] let assumption [ alb ] hold . then
, the function @xmath129 satisfies properties ( a ) and ( b ) stated in corollary [ lemmab ] . in addition , property ( a ) stated in corollary [ lemmab ] holds for all @xmath381 in view of lemma [ lem : fkde - sol ] , the function @xmath191 satisfies properties ( i ) and ( ii ) stated in theorem [ thm : fkde ] . in particular , it satisfies these properties for the smaller class of @xmath53-bounded sets .
thus , it follows from corollary [ lemmab ] that the function @xmath191 satisfies properties ( a ) and ( b ) stated in corollary [ lemmab ] . in addition , lemma [ lem : fkde - sol](a ) implies that property ( a ) stated in corollary [ lemmab ] holds for all @xmath381 the following corollary generalizes ( * ? ? ?
* theorem 4.3 ) .
the difference is that corollary [ corr2 ] states that @xmath129 is the minimal solution within the class of functions satisfying the weakly continuity property , when @xmath23 is a @xmath53-bounded set , while ( * ? ? ?
* theorem 4.3 ) claims the minimality within the smaller class of functions satisfying the weakly continuity property when @xmath381 [ corr2 ] ( cp .
feinberg et al .
* theorem 4.3 ) ) let assumption [ lb ] hold .
then @xmath129 is the minimal function in @xmath224 satisfying properties ( a ) and ( b ) stated in corollary [ lemmab ] .
furthermore , if the transition function @xmath191 is regular , then @xmath191 is the unique element of @xmath224 taking values in @xmath19 $ ] and satisfying properties ( a ) and ( b ) stated in corollary [ lemmab ] .
\(b ) the function @xmath191 is the minimal function in @xmath224 for which statement ( a ) holds .
in addition , if the transition function @xmath191 is regular , then @xmath191 is the unique function in @xmath224 with values in @xmath19 $ ] for which statement ( a ) holds .
lemma [ cor ] and corollary [ lemmab ] imply that statement ( a ) of the corollary holds for a function @xmath28 from @xmath382 if and only if the function @xmath28 satisfies properties ( a ) and ( b ) stated in corollary [ lemmab ] .
therefore , this corollary follows from corollary [ corr2 ] . when @xmath383 is fixed and @xmath384 , formula is an equation in two variables @xmath39 and @xmath23 .
hence , for simplicity , we write @xmath385 instead of @xmath386 in for any function @xmath28 from @xmath224 when @xmath383 is fixed and @xmath387 and becomes @xmath388 for fixed @xmath42 and @xmath384 , the function @xmath389 is the marginal probability distribution of the process @xmath30 at time @xmath39 given @xmath390 . under assumption [ lb ]
, the following corollary describes the minimal solution of and provides a sufficient condition for its uniqueness .
\(b ) @xmath392 where @xmath172t_0,t_1[$ ] and @xmath393 is the minimal non - negative function that is a measure on @xmath184 for fixed @xmath39 , is measurable in @xmath39 for fixed @xmath23 , and for which statement ( a ) holds .
in addition , if the function @xmath394 is bounded on the set @xmath395 , then @xmath391 is the unique non - negative function with values in @xmath19 $ ] and satisfying the conditions stated in the first sentence of this statement .
statement ( a ) of the corollary follows immediately from corollary [ cor - s](a ) when @xmath384 . to prove statement ( b ) ,
consider a non - negative function @xmath385 , where @xmath16t_0,t_1[$ ] and @xmath18 that satisfies the conditions given in the first sentence of statement ( b ) of this corollary .
define the function @xmath396 , @xmath397 then , it follows from corollary [ cor - s](a ) and that the function @xmath8 satisfies the property given in corollary [ cor - s](a ) .
thus , corollary [ cor - s](b ) and imply @xmath398t_0,t_1 [ , b \in { \mathfrak{b}}({\bf x}).\ ] ] to show the uniqueness property , let the function @xmath28 take values in @xmath19 $ ] .
this fact and the property that the function @xmath399 takes values in @xmath19 $ ] for all @xmath221 @xmath400 @xmath111 @xmath23 in the domain of @xmath191 imply that the function @xmath8 defined in takes values in @xmath19 $ ] .
observe that @xmath31 is a @xmath53-bounded set if the function @xmath394 is bounded on the set @xmath401 then , as follows from corollary [ cor - s](a ) , @xmath402 for all @xmath36 @xmath16u , t_1[,$ ] and @xmath403 .
therefore , it follows from corollary [ cor - s](b ) that @xmath404 for all @xmath405 in the domain of @xmath191 , which along with implies the uniqueness property of @xmath406 anderson , w.j .
_ continuous - time markov chains : an applications - oriented approach_. new york : springer - verlag .
bertsekas , d.p . , & shreve , s.e .
_ stochastic optimal control : the discrete - time case_. new york : academic press .
feinberg , e. a. , mandava , m. , & shiryaev , a. n. ( 2013 ) .
sufficiency of markov policies for continuous - time markov decision processes and solutions to kolmogorov s forward equation for jump markov processes . in _ proceedings of 2013 ieee
52nd annual conference on decision and control _
. 57285732 ) .
kolmogorov , a. n. ( 1992 ) . on analytic methods in probability theory ( in russian 1938 ) .
shiryaev ( ed . ) , _ selected works of a.n .
kolmogorov _ , vol .
ii , probability theory and mathematical statistics . springer .
royden , h. l. ( 1988 ) .
_ real analysis _ , 3rd edn .
new york : macmillan .
ye , l. , guo , x. , & hernndez - lerma , o. ( 2008 ) .
existence and regularity of a nonhomogeneous transition matrix under measurability conditions .
_ j. theoret .
_ , 21 ( 3 ) , 604627 . | as is well - known , transition probabilities of jump markov processes satisfy kolmogorov s backward and forward equations . in the seminal 1940 paper , william feller investigated solutions of kolmogorov s equations for jump markov processes .
recently the authors solved the problem studied by feller and showed that the minimal solution of kolmogorov s backward and forward equations is the transition probability of the corresponding jump markov process if the transition rate at each state is bounded .
this paper presents more general results . for kolmogorov s backward equation
, the sufficient condition for the described property of the minimal solution is that the transition rate at each state is locally integrable , and for kolmogorov s forward equation the corresponding sufficient condition is that the transition rate at each state is locally bounded . |
the lynds 1634 dark cloud is a small molecular core in the general proximity of the orion molecular cloud ( omc ) that shows a bright rim in h@xmath6 , indicating the presence of ionizing uv radiation .
core mass measurements are in the range of one or two dozen @xmath7 , e.g. , @xcite measured 12 @xmath7 in n@xmath5h@xmath8 and 28 @xmath7 in hco@xmath8 , while @xcite measured 3 - 10 @xmath7 .
the l1634 core contains at least three objects in the protostellar outflow phase .
the first outflow to be discovered was the system of herbig - haro objects 240 .
the brightest knot of this system of shock fronts is optically detectable as rno 40 and its proper motion away from the main embedded far infrared ( fir ) source later called iras 05173 - 0555 , but originally discovered by @xcite in a generally western direction was determined from photographic plates by @xcite .
the western outflow lobe is blueshifted , while the eastern lobe is redshifted @xcite .
the driving source of this spectacular outflow , the far - infrared source iras 05173 - 0555 , has a luminosity of 17 @xmath9 measured by @xcite .
there has been some debate as to whether the iras source should be classified as sed class 0 or 1 , summarized , e.g. , in @xcite .
its spectral energy distribution , sub - mm luminosity , and powerful outflow , as well as the infall signature observed at millimeter wavelengths would support classification as class 0 , while the fact that the central object is detectable at near infrared wavelengths argues for class i status . in any case , it is clear that iras 05173 - 0555 is a very young object with an estimated age of @xmath10 10@xmath11 yrs , still accreting substantially and driving a bright outflow with multiple shock fronts .
the near - infrared emission of the shock fronts associated with the l1634 iras 05173 - 0555 outflow was first seen in the @xmath12 survey by @xcite and then studied in detail in h@xmath5 10 s(1 ) line emission by @xcite and @xcite .
@xcite presented an imaging and spectroscopic study of the western half of the hh 240/241 outflow associated with iras 05173 - 0555 , concluding that the h@xmath5 10 s(1 ) emission is excited in internal c - shocks within the outflow . in their 1995 paper , hodapp & ladd identified a second , less luminous outflow source , labeled irs 7 , and two s(1 ) shock fronts plausibly associated with it that they labeled as objects 4 and 9 .
the fir map by @xcite had already indicated an extension of the emission to the east of the main fir source , and the authors at the time attributed this to emission from an hh object , but in retrospect , that extension indicated the presence of irs 7 . when compared to the iras 05173 - 0555 source , irs 7 is readily detectable in the near - infrared and thus less obscured , its sub - mm luminosity is lower , and its well - collimated outflow is less powerful .
all this indicates that irs 7 may be slightly more evolved than the fir iras source and may be a typical class i object in its late accretion phase . band , green is the s(1 ) line image , and red is the @xmath4 band .
the two outflow sources irs 7 and iras 05173 - 0555 are clearly visible by their redder color , while h@xmath5 shock fronts appear green .
the edge of the molecular cloud is faintly outlined by s(1 ) emission ( green ) , probably fluorescently excited .
the axes of the irs 7 outflow lobes are indicated . ] in their comprehensive study of the optical emission in l1634 @xcite identified optical [ sii ] emission knots along the same axis as defined by irs 4 , 7 , and 9 of @xcite and labeled these as components of the hh 979 system of herbig - haro objects . in fig . 1 , we present an overview color composite image of l1634 , based on our @xmath13 , s(1 ) narrow band , and @xmath4 images , with the objects found by @xcite and @xcite labeled and the axes of the irs 7 outflow indicated .
the next system of hh objects downwind to the nw of irs 4 was labeled hh 979 nw1 by @xcite . in the other outflow lobe ,
the @xcite hh 979 se1 system of s[ii ] emission knots is identical to the @xcite irs 9 system .
the fact that most shock fronts of the irs 7 outflow are optically detectable confirms again that this system is associated with less dense molecular material and is probably more evolved than iras 05173 - 0555 and its outflows .
more detailed sub - mm maps were published by @xcite and @xcite , clearly showing some sub - mm emission from the irs 7 source . while @xcite have mapped the outflow cavity of the main ( iras 05173 - 0555 ) flow , no nh@xmath14(1,1 ) emission was detected at the position of irs 7 .
@xcite identify a third outflow , labeled l1634n , which is located at the northern edge of the l1634 cloud . on our images , we do not detect any shock - excited h@xmath5 10 s(1 ) emission at that position .
@xcite have assumed a distance of 400 pc , on the basis of the best distance determinations to the orion nebula cluster of 389 pc based on vla parallax measurements by @xcite and 392 pc from the rotational properties of young stars by @xcite and the argument that the bright rim of l1634 is likely illuminated by the orion ob1 association stars
. this paper reports the discovery that the l1634 irs 7 outflow source is rapidly and periodically variable .
we measure a distance of l1634 of 404 pc , consistent with the arguments above .
we also present proper motion measurements of the h@xmath5 shock fronts over a time span of 19 years , providing convincing kinematic evidence that irs 7 is indeed the source of the outflow traced by irs 4 , irs 9 , and hh 979 .
as part of an on - going program of monitoring star - forming regions for variability , the area of l1634 was repeatedly imaged in the @xmath1 band with the iris 0.8 m telescope and 1k 2.5 @xmath15 m infrared camera @xcite of the universittssternwarte bochum on cerro armazones , operated jointly by the ruhr universitt bochum , germany and the universidad catolica del norte in antofagasta , chile . for each visit of a specific region
, iris takes 20 raw data frames on the object with an integration time of 20 s , with small dithers to get around detector artifacts , followed by 20 dithered data frames on a separate sky field chosen to avoid bright stars or extended emission . for these sky exposures ,
larger dithers are used than for the object exposures to minimize the impact of any extended emission .
after variability with a high amplitude and short time constants had been found in the near - infrared object irs 7 , we conducted a dedicated monitoring campaign , using the iris system and ukirt with wfcam @xcite as often as was practical in the observing seasons 2013 - 2014 and 2014 - 2015 .
the ukirt observation with wfcam were done with an individual integration time of 10 s , a dither pattern of 9 jitters , and separate sky exposures with the same parameters . in total , 27 ukirt images of good quality were obtained in the @xmath4 band .
two of the resulting ukirt images are shown in fig . 2 , and show irs 7 in the minimum phase and in the rising phase of the light curve .
the data from both iris and ukirt are shown in the light curve fig .
we have also observed l1634 irs 7 using wfcam on ukirt in the @xmath13 filter , and in the narrow - band s(1 ) line filter .
the @xmath13-band images were taken within minutes of the @xmath1 images and were therefore suitable for a measurement of the object color .
the s(1 ) line images formed the second epoch data set for a study of the shock - front proper motions . in the observing season 2014 to 2015 , we obtained spectra under gemini project gn-2014b - q-90 with the gemini near - infrared spectrograph ( gnirs ) @xcite in 5 nights , using a 0@xmath1645 wide slit centered on irs 7 .
the cross - dispersed mode was used , giving , in principle , the full spectrum from [email protected] @xmath15 m . however , in l1634 irs 7 , only the @xmath4-band emission was bright enough for spectroscopy .
the gnirs spectrograph slit was oriented at p.a .
145@xmath18 , aligned with the direction of the outflow of irs 7 given by the two emission knots irs 4 and 9 . in each of the 5 nights , six individual spectra of irs 7
were obtained , interleaved with six spectra of the night sky a few arcseconds away from irs 7 and any extended emission associated with it or the other shock fronts .
each spectrum consisted of six co - added exposures of 10 s integration time .
one additional spectrum was obtained on march 20 , 2015 , with the spex instrument @xcite at the infrared telescope facility ( irtf ) .
for these observations , a 0@xmath168 wide slit was used that covered much more of the extended s(1 ) line emission around irs 7 and was oriented at p.a . 160@xmath18
aligned with the outflow features close to the central object irs 7 .
the spex observations consisted of 11 on - object frames of 120 s integration time , interleaved with the same number of sky exposures . since these observations were taken with a wider slit and longer integration time than the gemini / gnirs spectra , the extended s(1 ) line emission contributes more to this spectrum and the s / n of the continuum is comparable , even though the irtf is the smaller telescope .
the iris camera raw data were processed in a reduction pipeline based on the image reduction and analysis facility ( iraf ) software @xcite .
the raw data were flatfielded using incandescent light dome flats .
sky frames were obtained at positions near the object specifically selected to be free of bright stars and extended emission .
the optical distortion correction was based on an as - built model of the optical system . after correcting the known optical distortions ,
the astrometric solutions of the individual images were calculated based on sextractor @xcite and scamp @xcite based on 2mass coordinates @xcite .
the images were then co - added using these astrometric solutions .
aperture photometry was obtained with the iraf task phot with an aperture of @xmath19 pixels ( @xmath20 ) and calibrated against an ensemble of stars from the 2mass catalog .
these 2mass reference stars were selected for proper brightness , being isolated from other stars and from nebular emission . for part of the monitoring period ,
the iris telescope experienced unusually large pointing errors and unstable image quality due to a mirror support problem , so that there are only 9 suitable 2mass photometric reference stars common to all iris frames used here .
the 2mass reference stars were in the brightness range 12.5 - 14.5 . for the iris data the rms scatter of the measured magnitudes ranged from 0.03 for the brightest to 0.10 mag for the faintest reference stars , while for the ukirt data , the rms variations were 0.007 for the brightest and 0.026 for the faintest reference stars .
these rms uncertainties are consistent with those measured on larger samples of reference stars in other fields , and do not indicate any significant variability of any of the reference stars . in fig . 3 , we show the measured data of three stars in the field , two reference stars and one other , bright , apparently constant star that was not used as a reference star because its image overlaps with a background galaxy labeled vla1 in @xcite , and its photometry quality flags therefore did not meet our criteria .
the brighter of the reference stars has a magnitude standard deviation of 0.007 for ukirt and 0.027 mag for iris , the fainter one has a magnitude rms of 0.013 for ukirt and 0.040 for iris .
the ukirt data were processed by the cambridge data reduction pipeline .
we used the data in the form of fully reduced and co - added images , and extracted the photometry from those images using the same procedures and reference 2mass stars as were used for the iris data . for consistency
, we also used the same large photometric aperture as for the iris data , even though the image quality of the ukirt data would have allowed much smaller apertures to be used . for the study of the relatively bright irs 7 , the resulting increase in background noise was irrelevant .
the ukirt wfcam uses a k - band filter based on the mauna kea observatory ( mko ) standard described by @xcite , and the color transformation between this wfcam @xmath4-band and the 2mass photometry has been given by @xcite .
our 2mass reference stars have an average @xmath21 color of 1.0 mag , while for irs 7 , the 2mass catalog gives @xmath21 = 4.5 mag .
the small but significant color correction of 0.035 mag has been applied to the ukirt photometry of irs 7 , so that all photometry in figs .
3 and 5 is in the 2mass @xmath1 system .
however , this subtle correction has no bearing on any of our conclusions . from all the @xmath1 photometric data over the time span from 2012 to 2015
, obtained either with iris or ukirt , a period of the variability of @xmath0 days was determined ( fig . 4 ) , using the lomb - scargle algorithm developed by @xcite and @xcite , and implemented in the period analysis software ( peranso ) written by t. vanmunster .
in addition to the lomb - scargle algorithm , peranso offers twelve other period determination algorithms .
we used all of these on our data set , and used the scatter of these independent period determinations as a measure of the error associated with the lomb - scargle period value . from this method
we estimate that the period of l1634 irs7 has an rms error of 0.04 days , i.e. , approximately one hour .
the light - curve resulting from folding all our data with the 37.14 day period is displayed in fig .
5 . with the maximum light at phase 0.5 .
l1634 spends about half the time , from phase 0.8 to 0.3 , near its minimum brightness , the other half of the period is an outburst with a brightness increase between one and two magnitudes in the @xmath1 band .
the brightness increase starts at phase 0.30 and the maximum is at phase 0.50 .
the declining phase of the light curve can roughly be divided into a fast decline , from phase 0.5 to about 0.8 , and then a slow , residual decline from phase 0.8 to approximately , and coincidentially , phase 1.0 .
a kolmogorov - smirnov comparison of the data in the phase range 0.8 to 1.0 to the data in the quiescent phase from 0.0 to 0.3 gave a 4@xmath22 confidence that these are not randomly drawn from the same population . from the best sampled observed maximum , the highest of all photometric values obtained was at mjd 57079.1145 with a @xmath1 = 11.23 magnitude .
in the phased light curve ( fig .
5 ) , this data point was at phase 0.4988 .
we estimate an uncertainty of the timing of this maximum of 0.1 d. from this we derive the following ephemeris for the maxima of l1634 irs 7 : mjd @xmath23 .
the photometric errors were determined from the scatter of the measurements on constant stars in the field around l1634 irs 7 , and are indicated as error bars in figs . 3 and 5 . these photometric errors , 0.01 mag rms for ukirt and 0.03 mag for iris at the magnitude level of the outburst , are much smaller than the variations observed in the light curve .
we conclude that the quiescent - phase brightness of l1634 irs 7 varies within a range of @xmath10 0.5 mag , and the peak brightness varies by as much as a full magnitude from outburst to outburst . in the last observing season 2014 - 2015 ( blue points in fig .
5 ) , the well - sampled light curve lies above all previously recorded data points , both in the quiescent phases and during the outbursts . by comparison
, in the 2013 - 2014 season ( green points in fig . 5 ) , the iris and ukirt data points were below most other data points .
this indicates variations in outburst amplitude , but the data from the earlier years do not indicate a secular brightnening trend . at every epoch
the wfcam data in the three filters used ( @xmath13 , s(1 ) , and @xmath4 ) were obtained within minutes of each other , and are therefore suitable for a direct comparison of the colors .
the @xmath24 color of irs 7 is plotted against the phase of the periodic variability in fig . 6 , and the average colors in the two major phase ranges are indicated .
we have divided the color data into two samples : color during the quiescent phase from 0.8 to 0.3 ( including the slow decline phase ) , and during the bright phase from 0.3 to 0.8 phase .
the two - sample kolmogorov - smirnov test gives only a 2.5% probability of these two samples being drawn from the same population . with a marginal level of significance , the @xmath24 color is redder during the bright phase .
the ukirt data were precisely aligned using the iraf task xregister and a small sub - image containing irs 7 was selected .
we then identified the light curve phase of all ukirt images and after further selecting only ukirt images with good seeing and free of other defects , those selected images were arranged in the order of their phase .
images with similar phases were added together for better signal - to - noise and form the basis for fig . 7 . in this figure
, we show the @xmath4-band images directly , but also have divided the individual phase images by the average of all low - light images , to normalize the scattering properties of the dust distribution around the object , and to bring out the spherical nature of the light echoes .
the scattering effects of a light echo can occur in light paths of any inclination angle @xmath25 against the line of sight , and for such light paths the measured projected light echo expansion velocity will be @xmath26 . in our data ,
the brightest individual light echo is seen along what appears to be the wall of the northern , blueshifted , outflow cavity , and this particular light echo is clealy moving slower than the projected speed of light , and since it corresponds to the blueshifted outflow lobe , is is clear that the scattering is at an angle of less than 90@xmath27 to the line of sight . for a distance measurement
, we have to consider the maximum projected velocity in the overall light echo , and work under the assumption that this maximum velocity corresponds to scattering under a 90@xmath27 angle can .
the molecular core l1634 clearly contains dust , and the young stellar object irs 7 has formed and is still located at a density enhancement of this dusty material .
there may be extinction effects along different scattering paths .
our method of normalizing the flux distribution at any given phase of the variation by the average of the low - light phase is designed to normalize out effects of uneven illumination or extinction in different scattering paths .
the result of this normalization are light echo images of nearly round appearance , within the limits of the poor signal - to - noise ratio of those normalized images .
the approximately round outer boundary of these images is therefore assumed to be light scattered at 90@xmath27 angle , i.e. , in the plane of the sky , since nothing in our data suggests a dust distribution contrary to this assumption .
the fact that the distance so determined is in good agreement with other , independent , measurements gives added confidence that our assumption is correct .
we have measured the expansion of the light echoes in the normalized light echo images by visually fitting a circle around the perimeter of the normalized light echo .
the diameters measured in this rather crude way are shown as caliper - symbols next to the normalized images in fig . 7 , to demonstrate that their uncertainties do indeed allow a measurement of the light echo expansion .
the measurements are plotted in fig . 8 against the phase of the light curve together with a linear regression fit and the associate confidence interval of the resulting distance .
if e is the expansion of 15@xmath16915 per full period of p = 37.14 d , the distance in km is : @xmath28 . from the confidence interval of the regression fit ,
the errors are @xmath2 35 pc . a straight line with a slope corresponding to the speed of light projected onto the sky at this distance of 404 pc is shown in fig . 8 .
this distance measurement is consistent with the distances to the orion nebula by @xcite and @xcite , the arguments presented by @xcite why l1634 must be physically close to the orion nebula , and with the distance of 400 pc assumed in this paper and by @xcite .
our measurements are of similar precision to those other distance determinations , and are entirely independent in their methodology , giving added confidence in this new consensus distance to the orion star forming region .
the spectra obtained in queue mode with gnirs on the fred gillette gemini north telescope were processed using standard procedures in iraf .
the raw data were dark subtracted , filtered to reduce pattern noise , and flat - field corrected .
the wavelength calibration is based on argon lamp spectra .
the @xmath4-band spectra were extracted using the iraf routine appall , and the resulting spectra were corrected for telluric absorption by division of an a0 star spectrum at very nearly identical air mass .
the strong br@xmath29 absorption line in the spectra of these calibration stars was replaced by a linearly interpolated continuum , so that we avoid creating artifacts at the nominal position of the br@xmath29 line .
finally , the spectrum was multiplied by an artificial spectrum of a 9700 k blackbody @xcite , to arrive at relative flux units .
similar procedures were used for the irtf / spex data , except that those spectra show more s(1 ) emission due to the wider slit .
the phase of the spectra was determined from their effective epochs .
it turned out that we got only one gnirs spectrum near maximum light , three gnirs spectra near a phase of 0.90 in the quiescent phase , and the spex spectrum at a phase of 0.10 , also in the quiescent phase , as shown by the arrow symbols in fig .
the three quiescent - phase gnirs spectra were co - added to improve the signal - to - noise ratio .
all spectral features appear to scale with the overall broad - band brightness , with the exception of the s(1 ) line that appears independent of the variations in irs 7 . for fig .
9 , the individual spectra were scaled to match their continuum fluxes , and then linearly shifted to separate them out for the display in the figure .
the shift vector is shown in fig .
8 as an arrow .
we computed the equivalent width of the @xmath30 ( 2 - 0 ) bandhead , the first and best isolated of the co bandheads .
the continuum was defined as in the co - index definition by @xcite as the range from 2.2873 - 2.2925 mum , and the line flux was integrated over the line wavelength range from 2.2931 - 2.2983 @xmath15 m .
the errors were computed from the standard deviation of the flux measurements in the continuum wavelength range .
the resulting equivalent width values are indicated in fig .
9 . the equivalent widths measured in the quiescent phases are higher than that measured during the outburst phase , but this effect is not significant at the 3@xmath22 level .
what can be stated is that the co - bandhead emission is not constant during the outburst , but it appears not to increase as strongly as the continuum flux .
emission of the co bandheads indicates an optically thin warm , dense inner disk around irs 7 .
the s(1 ) line emission does not scale with the brightness but is constant and independent of the continuum flux .
we conclude that the s(1 ) emission is generated sufficiently far away from irs 7 that the variations do not affect the emission of this line .
we have studied the proper motions of individual shock - excited emission knots in the outflows of l1634 iras 05173 - 0555 ( the main outflow ) and of irs 7 .
the first epoch data were data obtained in 1994 dec . 8 and 9 with a s(1 )
line filter in the quirc infrared camera @xcite on the uh 2.2 m telescope .
these data had originally been published in @xcite . for the purpose of this proper motion study ,
the data were re - reduced using modern methods .
we created separate co - added images of the western , center , and eastern parts of the l1634 outflow , each aligned using one reference star common to all the frames in that part of the outflow .
the effective epoch of the resulting images was mjd 49695.0 the second epoch data are the combined images in the h@xmath5 10 s(1 ) filter taken between 2013 nov .
7 ( mjd 56603.5 ) and 2014 april 16 ( mjd 56760.2 ) with ukirt / wfcam
. a total of 27 images were taken over this time span , and the effective epoch of the co - added image is mjd 56570.2 . the co - added first epoch sub - field images were aligned relative to the combined second epoch ukirt images with stars recorded in both images and by using the iraf geomap and geotran tasks .
this produced a version of the smaller first epoch image astrometrically matched to the system of astrometric reference stars in the larger ukirt second epoch image . in the process ,
outliers , i.e. , stars with high proper motions , were identified and eliminated .
after this prelimary alignment step , retangular areas were defined covering individual s(1 ) shock fronts , and the iraf task xregister was used to compute the spatial cross - correlation of these rectangular image sections at both epochs , giving the x and y shift required for maximum correlation .
this shift is the proper motion in pixels .
the rectangular areas used are indicated in fig .
the same process was used on stars in the image .
the proper motions measured on stars are generally much smaller than those in the shock fronts , and give information about the combined uncertainties caused by reference star proper motion and the typical errors of the cross - correlation maxima .
ccccc iras & 05173 - 0555 & outflow & + 05 19 35.9 & @xmath175 50 59 & 5 & 23 & hh240e ( new ) + 05 19 36.8 & @xmath175 51 15 & 25 & 20 & hh240d + 05 19 37.3 & @xmath175 51 22 & 17 & 15 & hh240c + 05 19 38.3 & @xmath175 51 30 & -2 & 12 & hh240b1
+ 05 19 39.1 & @xmath175 51 24 & 12 & 17 & hh240b2 + 05 19 39.2 & @xmath175 51 37 & 10 & 24 & hh240b ( se ) + 05 19 40.7 & @xmath175 51 43 & 35 & 22 & hh240a ( sw ) + 05 19 40.8 & @xmath175 51 32 & 45 & 18 & hh240a ( nw ) + 05 19 42.5 & @xmath175 51 58 & 17 & 7 & hh240a ( se ) + 05 19 42.7 & @xmath175 51 38 & 28 & 13 & hh240a ( ne ) + 05 19 44.1 & @xmath175 51 50 & 43 & 15 & hh240z ( new ) + 05 19 47.9 & @xmath175 52 03 & 48 & -2 & hh240y ( new ) + 05 19 49.6 & @xmath175 52 04 & -49 & -4 & hh241v ( new ) + 05 19 54.3 & @xmath175 52 25 & -8 & -8 & hh241z ( new ) + 05 19 56.1 & @xmath175 52 27 & -31 & -10 & hh241a + 05 19 58.6 & @xmath175 52 26 & -13 & -12 & hh241b + 05 20 01.4 & @xmath175 52 37 & -41 & -10 & hh241d + 05 20 03.5 & -05 52 33 & -17 & -3 & hh241e ( new ) + & & & + irs 7 & & outflow & + 05 19 46.7 & @xmath175 50 53 & 40 & 43 & irs4 ( w ) + 05 19 47.2 & @xmath175 50 50 & 47 & 18 & irs4 ( e ) + 05 19 57.1 & @xmath175 53 18 & -31 & -39 & irs9 = hh979se1 + the two epochs are separated by 6975 days or 19.097 years . in fig .
10 the measured proper motion vectors were scaled by 1000/19.097 to represent the extrapolated proper motion in 1000 years .
table 1 lists the measured proper motion values in units of mas y@xmath31 and labels the individual shock fronts using the nomenclature of @xcite for the irs 7 outflow and of @xcite and @xcite for the iras 05173 - 0555 outflow .
in particular for the latter , the labels in the existing literature were not sufficient to uniquely identify the shock fronts studied here , so we have assigned additional labels to some shock fronts , and added descriptive sub - labels to others . in some cases
, we had to resort to adding letter from the end of the alphabet to denote shock fronts upstream from
@xmath32 . in fig .
10 , we show the measured proper motions of four stars near the l1634 outflows as very short arrows in orange color .
the rms uncertainty of the measured proper motions of these stars is 4@xmath168 per 1000 years or 9 km s@xmath31 in each coordinate .
the typical proper motions of shock fronts are a factor of 10 larger than this .
the reference star system is therefore not the limiting factor in the precision of our measurements .
proper motions of bright , well defined shock fronts are highly significant , while measurements on faint or poorly defined shock fronts are only marginally significant .
young stars are generally variable , in particular if high - precision observations with space - based telescopes are considered . for young stars sufficiently evolved to be optically detectable , the variable star t tauri is the prototypical example for the small , irregular variability observed in virtually all of them .
based on optical photometry of a large sample of t tauri stars , @xcite identified three types of variability : rotationally modulated cool star spots , variable accretion leading to changes in the veiling observed in some stars , and variable circumstellar obscuration .
variability surveys of star - forming regions at infrared wavelengths , primarily with the warm @xmath33 space telescope , have expanded our knowledge of young star variability to objects sufficiently young and embedded to only be detectable at infrared wavelengths .
the orion nebula region was monitored in the @xmath33 irac channels 1 and 2 by the young stellar object variability ( ysovar ) project , and @xcite have published examples of the lightcurves obtained in their month - long high - cadence monitoring campaign . the one star among their published light curves that has some resemblance to irs 7 is the class i object 2 m j053435.44 - 053959.1 , which they describe as being caused by slow changes of accretion over the 40 day time span of the observations . variability dominated by accretion bursts
was studied in ngc2264 by @xcite and @xcite using ysovar @xmath33 data .
however , these accretion burst stars are characterized by shorter , more frequent , and non - periodic outbursts .
the few examples in their sample with significant periodicity have periods in the range of 7 days .
it should be noted that their whole observing campaign of ngc2264 lasted only about 40 days , and a procedure for subtracting long - term trends in the light curves was used , so that this study has a reduced sensitivity to variations on timescales longer than @xmath10 10 days , and would not be able to recognize periodicity with periods of 37 days , such as irs 7 in l1634 . in summary , the results by @xcite and @xcite show that outburst phenomena are frequently observed in ysos on timescales shorter than those observed in irs 7 . on the other side , more substantial outbursts of long duration and larger amplitude in young , still accreting stars are traditionally classified into either fu orionis ( fuor ) or ex lupi ( exor ) type outbursts , depending on the duration of the outburst and its spectrum .
the first review of these phenomena has been given by @xcite .
fuor outbursts exhibit time scales of decades to centuries and the spectral characteristics of an optically thick luminous disk with absorption line spectra , including co bandhead absorption @xcite .
the less substantive exor outbursts have timescales of years , have been observed by @xcite to actually return to pre - outburst brightness , and show optically thin emission line spectra , in particular , co bandhead emission . in some cases , e.g. , v1647 ori , repetitive outburst have been observed @xcite .
fuor and exor outbursts are rare phenomena and most known cases have been discovered by optical methods where large surveys and monitoring campaigns had been feasible in the recent past .
a small number of more deeply embedded outburst objects have now been found that are only detectable at infrared wavelengths .
most of these can not be clearly classified as fuor or exor outbursts , but exhibit outburst amplitudes , durations , and spectral characteristics different from either of those two established classes .
one example is oo ser @xcite that showed an outburst amplitude of only 2.5 mag and a duration of the fast part of the outburst of only a few years , but has not settled back to its pre - outburst brightness in the two decades since its outburst .
an outburst of an even more deeply embedded object , hops 383 , a class 0 protostar , was recently reported by @xcite , showing that accretion instabilies can occur very early in the evolution of a protostar , when the object emits most of its energy at far - infrared and sub - mm wavelengths .
@xcite have recently discussed the various outburst phenomena in the context of determining the rate of fuor outburst , the highest amplitude and longest duration extreme of the distribution of outburst events . from this , and many papers discussing individual objects , it is clear that outburst events in young stars span timescales from hours to millennia , and burst amplitudes ranging from a few percent to several orders of magnitude in flux .
the amplitude and period of irs 7 fits right into this distribution , but the well - established periodicity makes it stand out among otherwise similar objects .
it is not unique in this respect , however .
for example , @xcite reported on the periodic ( 543d ) variability of the illuminating star v371 ser of the `` cometary '' nebula ec 53 in serpens nw , a very young object associated with an h@xmath5 jet and showing other signs of extreme youth . for irs 7 ,
the fact that light echoes can be observed to propagate into the surrounding dusty material ( fig . 7 ) with the speed of light at the distance of 404 pc argues against any effect that is specific to the line of sight towards the observer , e.g. , an eclipsing binary or periodic obscurations by a tilted disk or orbiting dust cloud . also , the near color neutrality of the variation ( fig . 6 ) speaks against a dust obscuration effect .
rotational modulation of an accretion hot spot at first seems a plausible explanation of the irs 7 variability , since it spends about have of the time in the bright state , and the other half at nearly constant , low light .
however , the light echoes appear round and are centered on the illuminating object . a hot spot on a rotating star or disk would be expected to form alternating light echoes on either side of the star , if the scattering is close to isotropic .
the only spectral feature clearly detected in our @xmath4-band spectra are the co bandheads in emission and emission from shock - excited h@xmath5 .
the shock excited s(1 ) line flux appears independent of the phase of the variability , while the co bandhead emission scales approximately with the overall continuum flux variations .
the very young evolutionary state of irs 7 as the driving source of a bipolar outflow suggests that accretion luminosity in a warm disk is the main source of luminosity .
the fact that the 2.29 @xmath15 m co bandheads are seen in emission indicates that the warm inner parts of the accretion disk are optically thin , and in this respect irs 7 resembles an ex lupi eruptive variable @xcite .
the co bandhead emission at 2.29 @xmath15 m is present both in the quiescent and the outburst phase , indicating that an optically thin warm disk is always present .
the equivalent width of the @xmath34 emission is slightly smaller in the outburst phase than in the quiescent phase , indicating the the co emitting part of the disk is not primarily responsible for the increase in luminosity during the outburst .
we caution that this result is statistically only marginally significant .
the co line emission is stronger during outburst , but not quite by the same factor as the continuum emission .
it is therefore reasonable to assume that the variability is caused by instabilities in the accretion process and that the warm gas responsible for the co bandhead emission never completely disappears , even during the low accretion phase of the light curve .
a short period binary companion is a possible explanation for the periodic accretion instabilities .
this mechanism of binary orbit triggered accretion instability was discussed by @xcite and modeled in detail by @xcite . while possible , this binary - triggered accretion model has the disadvantage of postulating the existence of a close binary companion , which we have no other evidence for . assuming a mass of 0.5 @xmath7for irs 7
, a 37 d = 0.10 y orbital period has a 0.11 au semi - major axis .
such a close binary orbiting within the disk feeding the accretion would , most likely , disrupt that disk in a short time .
thus we conclude that a short period binary companion is an unlikely explanation for the periodic accretion .
we prefer a model where the accretion instabilities are intrinsic to the accretion process itself , and do not require an external trigger .
a theoretical framework has recently been developed by @xcite for magnetosphere - disk interactions in the case of a star with strong dipolar magnetic field surrounded by a thin accretion disk in keplerian rotation .
they find that under certain conditions , accretion stalls out and matter accumulates in the disk just outside of the co - rotation radius .
when the gas pressure in that accumulated material becomes high enough , it overcomes the centrifugal barrier created by the interaction between the magnetic field and the disk , and accretion resumes .
after the material has been transferred to the star , the inner disk edge will move back outside the co - rotation radius , and the stage is set for a repetition of this process .
this mechanism provides a plausible explanation for cyclical accretion onto a young star . in a follow - on paper ,
@xcite specifically apply their model to the repetive eruptions of the prototypical exor ex lupi .
they make the point that in addition to the major outbursts , higher frequency oscillations with a period of about 30 d and amplitude of 1 2 mag are observed .
these higher frequency oscillations are very similar to the variations observed in l1634 irs 7 .
the comparison of the 1994 uh / quirc and 2014 ukirt / wfcam data clearly shows proper motion of the faint emission knots to the nw and se of irs 7 infrared sources 4 and 9 in the nomenclature of @xcite consistent with them originating from irs 7 .
the measured tangential velocities are 112.3 km s@xmath31 for irs 4 ( nw of irs 7 ) and 94.1 km s@xmath31 for irs 9 4 ( se of irs 7 ) . from the fact that we see near - infrared radiation , presumably scattered light , to the nw of irs 7 , but not in the opposite direction
, we conclude that the nw lobe of the bipolar outflow from irs 7 is oriented towards the observer , i.e. , blueshifted .
we detect the hh 979 nw1 shock front found by @xcite , but do not have first epoch s(1 ) images to determine a proper motion .
we do not detect any s(1 ) emission from hh 979 nw2 , where @xcite found another [ sii ] shock front .
this could be explained by this farthest shock front having broken out of the l1634 molecular cloud through its front surface .
we also measured the proper motions of many of the shock fronts and emission knots in the main outflow originating in iras 05173 - 0555 .
in particular , we measure proper motions in opposite direction for the two small shock fronts closest to the position of the iras source , proving that the main outflow does indeed originate from this deeply embedded protostar .
we detect scattered continuum light in the @xmath4 band ( red in figs . 1 and 10 ) from the inner walls of the outflow cavity of iras 05173 - 0555 .
the driving source itself , or the innermost parts of the outflow cavity in the direction of the blueshifted western outflow lobe , is seen as an unresolved continuum emission source in our wfcam @xmath4-band image , while it is undetected in the @xmath13 band and only marginally detected in the narrow - band s(1 ) image .
this continuum source is also located at the apex of the thin lines of continuum emission that apparently outline the walls of the outflow cavity in the west .
a similar outline of the eastern outflow cavity walls is also visible , but can not be traced all the way to the unresolved source , due to absorption from the nearly edge - on disk surrounding the driving source .
the point source , presumably at or very near the driving protostellar source of the outflow , is located at @xmath35 = 05:19:48.37 , @xmath36 = -05:52:03.9 , consistent with the positions given by @xcite for the sub - mm source , and by @xcite for the peak of the nh@xmath14 position in vla measurements .
the two innermost s(1 ) shock fronts in the iras 05173 - 0555 outflow are not equidistant from this position of the driving source .
the eastern shock front is 18@xmath37 east of the central source , while the western shock front is only 63 from that source .
the mid - point between these innermost shock fronts located at @xmath35 = 05:19:48.74 does not show any detectable @xmath4-band continuum source , is not morphologically consistent with the apex of the cavity outline , and does not coincide with any significant feature in the maps of @xcite and @xcite .
the projected velocity of the eastern shock is 94.3 km s@xmath31 while the western shock is 93.0 km s@xmath31 .
such high velocities would dissociate the h@xmath5 molecule if the shock were against stationary gas .
we conclude that both shocks are internal shocks with much lower relative shock velocity in a wind moving with @xmath10 90 km s@xmath31 . exactly where detectable s(1 ) emission is produced depends on the details of the internal velocity structure of the wind , and is plausibly different in the two lobes . @xcite measured radial velocities in both the blue - shifted ( hh240 ) and red - shifted ( hh241 ) lobe of the iras 05173 - 0555 outflow , relative to the systemic velocity of + 8 km s@xmath31 obtained by @xcite .
the radial velocities cover a broad range , but are generally centered around 30 km s@xmath31 .
compared to the tangential velocities derived here from the astrometry of order of 90 km s@xmath31 , this suggests that this outflow is oriented only about 18@xmath27 from the plane of the sky , consistent with the morphology .
while @xcite found the hh 980 system of herbig - haro objects associated with l1634n in their [ sii ] image , we do not detect any h@xmath5 10 s(1 ) emission from this yso .
there is some s(1 ) emission detectable in that general area , but this appears to be the fluorescently excited outer layer of the molecular cloud and associated with the photo - dissociation of h@xmath5 by the interstellar radiation field .
most of the l1634n outflow propagates outside of the bright rim of l1634 , in a region devoid of h@xmath5 .
the embedded outflow source irs 7 in the l1634 molecular cloud was found to experience periodic outbursts in brightness with a period of @xmath0 days and a typical peak - to - peak amplitude of 2 mag .
both the quiescent light brightness and the outburst amplitude are changing on the level of 0.5 mag over the monitoring period in a apparently random fashion with time constants many times larger than the main 37.14 day period .
the outbursts are associated with only small changes in the @xmath24 color , and we do not observe significant changes in the shape of the spectrum or individual spectral features .
the rapid outbursts produce light echoes propagating into the surrounding molecular cloud at the projected speed of light , demonstrating that the brightness variations of irs 7 are due to a mechanism that illuminates the surrounding dusty cloud isotropically .
the expansion of the light echoes allows the distance to l1634 to be measured as d = 404 pc @xmath2 35 pc .
we have used images in the s(1 ) line in comparison with similar data obtained in 1994 to measure the proper motions of the s(1 ) shock fronts in the two southern outflows .
this confirms that there are indeed two independent outflows , the larger one powered by the fir source iras 05173 - 0555 , the other by irs 7 .
most photometric data on l1634 irs 7 were obtained at the iris telescope of the universittssternwarte bochum on cerro armazones , which is operated under a cooperative agreement between the `` astronomisches
institut , ruhr universitt bochum '' , germany , the `` universidad catolica del norte '' in antofagasta , chile , and the institute for astronomy , university of hawaii , usa . the operation of the iris telescope is supported by the `` nordrhein - westflische akademie der wissenschaften und der knste '' in the framework of the academy program by the federal republic of germany and the state of nordrhein - westfalen .
construction of the iris infrared camera was supported by the national science foundation under grant ast07 - 04954 .
we wish to thank angie barr dominguez , thomas dembsky , holger drass , francisco p. nunez , lena kaderhandt , michael ramolla and christian westhues for operating the iris telescope for the acquisition of the data used in this paper , ramon watermann for writing the data reduction pipeline , and roland lemke for technical support .
the spectroscopy was mostly obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) .
one spectrum was obtained at the infrared telescope facility ( irtf ) , which is operated by the university of hawaii under contract nnh14ck55b with the national aeronautics and space administration .
we thank j. s. bus , m. connelley , and j. t. rayner for their support during the observations .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center/ california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
elias , j. h. , joyce , r. r. , liang , m. , muller , g. p. , hileman , e. a. , & george , j. r. 2006 , in ground - based and airborne instrumentation for astronomy , eds .
i. s. mclean and i. masanori , spie , 6269 , 138 skrutskie , m. f. , cutri , r. m. , stiening , r. , weinberg , m. d. , schneider , s. , carpenter , j. m. , beichman , c. , capps , r. , chester , t. , elias , j. , huchra , j. , liebert , j. , lonsdale , c. , monet , d. g. , price , s. , seitzer , p. , jarrett , t. , kirkpatrick , j. d. , gizis , j. , howard , e. , evans , t. , fowler , j. , fullmer , l. , hurt , r. , light , r. , kopan , e. l. , marsh , k. a. , mccallon , h. l. , tam , r. , van dyk , s. , & wheelock , s. 2006 , , 131 , 1163 | the small molecular cloud lynds 1634 contains at least three outflow sources .
we found one of these , irs 7 , to be variable with a period of @xmath0 days and an amplitude of approximately 2 mag in the @xmath1 band .
the light curve consists of a quiescent phase with little or no variation , and a rapid outburst phase . during the outburst phase
, the rapid brightness variation generates light echoes that propagate into the surrounding molecular cloud , allowing a measurement of the distance to irs 7 of 404 pc @xmath2 35 pc .
we observed only a marginally significant change in the @xmath3 color during the outburst phase .
the @xmath4-band spectrum of irs 7 shows co bandhead emission but its equivalent width does not change significantly with the phase of the light curve .
the h@xmath5 10 s(1 ) line emission does not follow the variability of the continuum flux .
we also used the imaging data for a proper motion study of the outflows originating from the irs 7 and the fir source iras 05173 - 0555 , and confirm that these are indeed distinct outflows . |
magnetic materials with face - centered cubic ( fcc ) lattice are characterized by a great variety of magnetic states , _
e.g. _ , antiferromagnets of type - i ( mnte@xmath1 @xcite ) ; type - ii ( mno @xcite ) ; type - iii ( mns@xmath1 @xcite ) ; spiral ( helical ) magnetic structures ( @xmath2-fe @xcite ) . there is a number of complex compounds where magnetic moments form an effective fcc lattice , _
e.g. _ , alkali fullerides a@xmath3c@xmath4 ( a = k , rb , cs ) @xcite , cluster compounds like gata@xmath5se@xmath6 , ganb@xmath5se@xmath6 @xcite , b site ordered double perovskites @xcite .
they demonstrate various magnetic orders as well and undergo metal - insulator transition ( mit ) under pressure .
the wealth of magnetic structures is largely explained by inherited geometric frustration of fcc lattice .
the lattice can be decomposed into the system of edge - sharing tetrahedra , which makes impossible the simple nel antiferromagnetic ( af ) order @xcite . the frustrated magnetism in fcc structure is traditionally studied in the framework of the heisenberg model which describes localized spins coupled by exchange interaction .
if nearest - neighbor exchange integral @xmath7 only is taken into account , the fourier transformation of the exchange interaction @xmath8 approaches its minimum at @xmath9 where @xmath10 is an arbitrary number in the @xmath11 $ ] range ( lattice constant is taken as a unit ) , which determines the type of the ground state magnetic order being unspecified .
for @xmath12 we have type - i af structure , for @xmath13 type - iii af structure , the intermediate range corresponding to spiral ( or spin - wave ) structures which are , generally speaking , non - collinear and incommensurate .
this ambiguity indicates the ground state degeneracy which is characteristic for frustrated systems .
finite temperatures lift the degeneracy by favoring collinear magnetic structures owing to so - called `` order from disorder '' effect @xcite .
the account of the next - nearest - neighbor exchange @xmath14 lifts the degeneracy already within the mean - field theory : depending on @xmath14 , one of the following collinear ground states is stable : type - i , type - ii ( @xmath15 ) , type - iii antiferromagnetic or ferromagnetic order @xcite .
carlo simulations predict type - i antiferromagnet at low temperatures for the nearest - neighbor model , while inclusion of @xmath14 should lead to type - iii antiferromagnet which goes to type - i with increasing temperature @xcite .
an account of magnetic anisotropy ( _ e.g. _ , dzyaloshinsky - moriya interaction ) can stabilize type - i or type - iii af order @xcite . in ref .
@xcite , monte
carlo simulations were performed for gdptbi with exchange constants determined from first - principle calculations , and type - iii antiferromagnet was found as the ground state .
the localized magnetism picture is insufficient to describe properties of itinerant fcc magnets .
in particular , the calculations based on the heisenberg model , as a rule , suppose exchange constants to be not changed with temperature , and can miss important effects : _
e. g. _ the results obtained for the hubbard model on the square lattice show a significant change of the spiral magnetic order parameters with temperature @xcite .
an alternative approach is provided by _
ab initio _ band structure calculations of specific fcc materials based on the density functional theory with possible extensions like lda+@xmath16 or lda+dmft which allow to take into account electron correlations .
this approach enables one to describe the experimentally observed ferromagnetic state in ni and to explain peculiarities of its electron spectrum caused by correlation effects @xcite .
modern band calculations reproduce successfully the experimental incommensurate spiral magnetic order in fcc @xmath2-fe revealing an important role of long - range competing exchange interactions @xcite . despite the considerable progress in predicting the properties of specific materials , band calculations lack generality and
often can not explain the physics leading to stabilization of one or another magnetic structure . to investigate the relative stability of all the possible commensurate and incommensurate magnetic states and to determine the role of correlations effects , depending on the system parameters ,
more general and fundamental approaches are needed .
the present paper is devoted to the construction of magnetic phase diagrams for the fcc lattice in the framework of hubbard model which contains all essential physics of metallic magnetism and metal - insulator transition . in order to study spiral magnetism and electron correlations ,
we use the slave - boson approach ( sba ) proposed by kotliar and ruckenstein @xcite and generalized to spiral magnetic ordering @xcite . in the saddle point approximation , sba is qualitatively similar to the gutzwiller approximation @xcite , the ground state energy being in a fair agreement with the quantum monte
carlo and exact diagonalization calculations @xcite . to evaluate the impact of electron correlations on the formation of magnetically ordered states and related metal - insulator transition we have used the hartree
fock approximation ( hfa ) as well .
the hfa method yields qualitatively correct estimates for the energy of magnetically ordered phases even for large @xmath16 , at least near half - filling @xcite . at the same time , in contrast to hfa , sba yields a correct interpolation description of paramagnetic phase energy too .
the sba method was used to construct magnetic phase diagrams for the square , simple cubic and bcc lattices @xcite , providing reasonable results in a description of electron correlations and essentially improving the generalized hfa @xcite .
recent hfa and sba were also applied to study magnetic and paramagnetic mit s for bipartite square , simple cubic and bcc lattices @xcite , importance of the spiral incommensurate magnetic states being demonstrated .
we treat the single - band hubbard model for the one - electron spectrum @xmath17 , @xmath18 where @xmath19 ( @xmath20 ) is a creation ( annihilation ) electron operator in one - electron state with quasimomentum @xmath21 and spin projection @xmath22 , and @xmath16 is coulomb ( hubbard ) interaction parameter , @xmath23 is particle number operator at site @xmath24 . to treat the spiral magnetic order formation
we apply the generalized sba @xcite and directly minimize the ground state _ grand potential _
@xmath25 with respect to spiral wave vector @xmath26 .
the calculations are based on the fermionic part of the effective hamiltonian , @xmath27 where @xmath28 @xmath29 @xmath30 is the correlation - induced one - electron band narrowing and @xmath31 is the one - electron shift .
the average probability amplitudes of quantum many - electron states ( empty @xmath32 , singly occupied @xmath33 , doubly occupied @xmath34 ) are calculated in a mean - field manner .
the corresponding eigenvalues ( antiferromagnetic subbands ) yielding the renormalization of the electron spectrum can be expressed analytically , @xmath35 where @xmath36 the bare electron spectrum in fcc lattice with account of nearest- and next - nearest neighbors hopping reads @xmath37 the electronic concentration @xmath38 and the magnetization @xmath39 satisfy mean - field - like equations @xcite respectively , @xmath40 @xmath41 being in contrast with the former equations by that renormalization parameters @xmath42 and @xmath31 are directly taken into account .
the minimization of @xmath25 with respect to @xmath26 was performed numerically with @xmath43 running in high - symmetry directions of the brillouin zone : @xmath44 , @xmath45 , @xmath46 , @xmath47 , @xmath48 and @xmath49 directions were taken into account .
the step of changing @xmath50 was @xmath51 .
the integration over the brillouin zone was performed on the grid containing 70 @xmath21-points in each direction .
the accuracy settings were adjusted , if necessary . since the ground state @xmath52 actually depends on the chemical potential @xmath53 as a parameter
, we can determine the dependence of the magnetic structure on @xmath53 taking into account the possibility of the phase separation automatically @xcite , avoiding the tedious application of the maxwell rule in the negative compressibility regions .
sba reproduces well - known brinkman - rice ( br ) result @xcite for paramagnetic - state mott transition .
the consideration of magnetic states leads to a possibility of mit by slater scenario : a rise of @xmath16 increases the magnetic moment and opens a gap in the electron spectrum at some critical @xmath54 .
for bipartite lattices with only nearest - neighbor electron hopping the magnetic mit occurs at @xmath55 , magnetic insulator state being nel antiferromagnet .
an account of next - nearest - neighbor hopping @xmath56 results in finite values of @xmath54 . in order to study the metal - insulator transition we fix the electron concentration @xmath57 adjusting @xmath53
correspondingly . in the insulating
magnetically ordered state the chemical potential should be placed in the gap of one - electron spectrum @xmath58 , see eq .
( [ eq : subband_spectrum ] ) , @xmath59 this implies that in our model any deviation from half - filling results in the shift of the fermi level from the gap into one of subbands and , therefore , in impossibility of insulator state . to calculate the critical value of mit @xmath54 we run over values of @xmath16 starting from zero until mit is found . in the case of second - order magnetic phase transition
the determination of @xmath60 is not sufficiently precise ( we assume @xmath61 ) , and we determine it by calculating the generalized magnetic susceptibility within sba @xcite . while type - i and type - ii af structures can be viewed as spiral magnetic orders with @xmath62 and @xmath63 correspondingly , for the type - iii af structure this is not possible .
@xmath64 wave vector , which is associated with type - iii antiferromagnet , gives actually a _ non - collinear _ structure as a regular spiral depicted in fig .
[ fig : af3]a .
it can be divided into two sublattices ( a and b ) , each being ordered antiferromagnetically ( filled and empty arrows in fig . [
fig : af3 ] ) , so that their magnetic moments are perpendicular to each other . to obtain the type - iii collinear order the magnetization direction of sublattice b should be rotated by @xmath65 ( fig .
[ fig : af3]b )
. this structure can be described by @xmath66 as well , but its site magnetization dependence is specified by @xmath67 , @xmath68 ( that is a commensurate spin density wave ) , in contrast to spiral s one @xmath69 , @xmath70 .
as for experimental observations of the type - iii af phase ( see introduction ) , a collinear structure is usually implied . experimental distinction between collinear type - iii af and non - collinear @xmath64 orders is a non - trivial task which requires some special techniques such as polarized neutron diffraction on single crystals .
the application of these measurements to the compound mns@xmath1 confirmed the formation of collinear type - iii af order at low temperatures @xcite .
spiral structure .
b ) collinear type - iii af structure . filled and empty arrows belong to the antiferromagnetic sublattices a and b respectively . , scaledwidth=49.0% ] the localized - moment picture predicts the same energy for the collinear type - iii af and non - collinear @xmath64 states . this can be easily understood : total intersublattice interaction energy equals to zero due to antiferromagnetic structure and intrasublattice interaction energy does not depend on the relative orientation of sublattices . in the itinerant - electron systems these two states
are not generally energetically degenerate and the calculations within the hubbard model can identify the ground state ; such an investigation was not performed yet to our knowledge . for the sake of generality , we assume that the sublattice b magnetization is rotated by some angle @xmath71 as compared to the trivial spiral order .
analogously to the eqs .
( [ hf]),([eq : t ] ) we write the fermionic part of the hubbard hamiltonian for the system of two sublattices , @xmath72 where @xmath73 is a sublattice index ; intrasublattice ( s ) and intersublattice ( i ) fourier transforms of the hopping read @xmath74 so @xmath75 and @xmath76 correspond to the spiral @xmath64 and the collinear af - iii structures respectively .
the bare electron spectra within ( s ) and between ( i ) sublattices have a form respectively , @xmath77 diagonalization of the hamiltonian ( [ eq : ham_phi ] ) with double number of @xmath78-operators can not be done analytically and requires numerical calculation of spectra in each @xmath21 point .
performing this procedure allows us to calculate the ground state thermodynamical potential @xmath52 and include the collinear af - iii phase on the equal foot with a full set of spiral phases for the construction of the hubbard model phase diagrams . within sba .
the spiral phase regions are denoted according to the form of their wave vector ( concrete _ number _ @xmath50 depends on the point @xmath79 of the region ) .
filling shows the phase separation regions .
bold ( blue ) lines denote the second - order phase transitions .
solid ( red ) lines correspond to the boundaries between the regions of the homogeneous phase and phase separation .
bold dashed ( red ) lines denote the first order phase transitions in the case where the region of the phase separation is narrow . dashed ( red ) horizontal lines
separate the phase separation regions corresponding to different phase pairs .
, scaledwidth=49.0% ] .
the notations are the same as in fig .
[ fig : fcc_t=0 ] .
, scaledwidth=49.0% ] .
the notations are the same as in fig .
[ fig : fcc_t=0 ] .
, scaledwidth=49.0% ]
we calculated the sba phase diagrams of the ground state at following values of @xmath0 : @xmath80 ( fig . [
fig : fcc_t=0 ] ) , @xmath81 ( fig .
[ fig : fcc_t=+0.3 ] ) and @xmath82 ( fig .
[ fig : fcc_t=-0.3 ] ) . the diagrams are rich and contain the regions of paramagnetic , commensurate and incommensurate magnetic states .
the commensurate phases include all three types of antiferromagnetic structures observed in real materials .
the region of @xmath49 incommensurate spin - spiral order , observed in fcc - fe @xcite , is presented for all @xmath0-values , being especially large for @xmath83 .
first - order transitions on the diagrams are accompanied by the phase separation regions .
similar diagrams and their detailed analysis were presented in ref .
@xcite , but without including the type - iii af state . here
we focus on the differences owing to this additional phase .
we see that for all considered values of @xmath0 the af - iii phase becomes the ground state at half - filling and in its vicinity , replacing the non - collinear @xmath64 phase ( see figs .
10 - 12 in ref .
@xcite for comparison ) .
the only case where the @xmath64-order survives is @xmath84 one where the narrow region of the non - collinear state is present at @xmath85 and @xmath86 ( fig .
[ fig : fcc_t=-0.3 ] ) .
the wide area of the phase separation between @xmath64 spiral state and type - iii af phase for @xmath84 suggests a possibility of an intermediate magnetic order formation . while af - iii structure results from @xmath64 order by rotating the af sublattice b by @xmath87 ( fig .
[ fig : af3 ] ) , the general intermediate structure can be specified by an arbitrary angle @xmath71 between sublattices , see eq .
( [ eq : ham_phi ] ) .
we choose a point @xmath88 , @xmath89 where the energies of pure @xmath64 and af - iii phases are almost equal .
for these parameters the total energy dependence on the angle @xmath71 between af sublattices is calculated ( fig .
[ fig : phi ] ) .
the graph has two clear minima at @xmath80 and @xmath65 angles , so the intermediate angles turn out to be energetically unfavorable .
an analogous behavior was found for other parameters studied .
note that extreme energy values differ on the fourth decimal place in the units of @xmath90 .
this means that all these magnetic states can be hardly distinguished between each other by limited in accuracy numerical simulation methods , like monte carlo .
even if the required accuracy is reached the potential barrier between @xmath64 and af - iii states does not allow to obtain @xmath64 phase starting from the af - iii state and vice versa during simulation time .
corresponding to @xmath75 spiral order .
the energy of type - iii af state ( @xmath76 ) is almost equal and differs in 7-th decimal place .
, scaledwidth=49.0% ] in ref .
@xcite the hubbard model on fcc lattice was studied in the dynamical mean - field theory taking into account the commensurate phases only .
a finite temperature quantum monte - carlo method was used to calculate the phase diagrams extrapolated to @xmath91 .
no fm order was found for @xmath92 and @xmath93 , which is in accordance with our diagram ( fig .
[ fig : fcc_t=0 ] ) where the fm region for small values of @xmath94 is very narrow and can be hardly detected in finite temperature calculations . in the case
@xmath95 no magnetic order was found for @xmath96 and the sequence of magnetic states af@xmath97pm@xmath97fm@xmath97pm was found for @xmath98 starting from half - filling .
this perfectly agrees with our diagram for @xmath84 ( fig .
[ fig : fcc_t=-0.3 ] ) except for the layer of the spiral @xmath47-state between af and fm regions ( instead of pm layer ) , which was not considered in ref .
this qualitative agreement confirms the validity of the results obtained in this section . to estimate the effect of electron correlations we present the analogous magnetic phase diagram at general @xmath38 for @xmath92 using hfa ( fig .
[ fig : fcc_t=0_hfa ] ) .
( the particular case of mit ( @xmath99 ) is discussed below , see fig .
[ fig : mit_hf ] . )
one can see that the hfa diagram turns out to be considerably more rich and complicated than the sba one , especially for @xmath100 .
thus correlations result in reducing the variety of spiral magnetic phases .
we have strong renormalization of fm phase boundaries at @xmath101 by the cost of extension along @xmath16 axis of small @xmath16 spiral phase regions . on the other hand , at @xmath102 the correct description of paramagnetic phase energy in sba results in suppression of spiral magnetic phases lying in the vicinity of pm phase region boundary . in hartree
fock approximation .
the notations are the same as in fig .
[ fig : fcc_t=0 ] .
, scaledwidth=49.0% ]
an interesting feature of the diagrams [ fig : fcc_t=0][fig : fcc_t=-0.3 ] is the metal - insulator transition which occurs in the half - filled band ( @xmath57 ) due to opening the gap in electron spectrum with increasing @xmath16 .
for all three @xmath56 values considered the insulator state appears to be the type - iii af one , the critical values @xmath54 of mit being approximately @xmath103 , @xmath104 , @xmath105 for @xmath92 , @xmath106 and @xmath107 correspondingly . for a more detailed study of this important phenomenon
we construct phase diagram of the model for @xmath57 in terms of @xmath16 and @xmath56 variables .
[ fig : mit ] presents the phase diagram of the ground state at half - filling for @xmath108 with account of all the spiral states and the collinear type - iii af phase .
we especially emphasize the important case of nearest - neighbor approximation ( @xmath92 ) , discussed earlier in the heisenberg model .
while the inclusion of nonzero electron hopping integral @xmath90 ( band picture ) makes the collinear type - iii af order to be favorable among other antiferromagnetic structures , the energy difference between type - i , collinear type - iii and non - collinear ( @xmath109 ) orders tends to zero as @xmath90 decreases ( @xmath110 ) . and @xmath0 in sba .
solid ( blue ) lines denote the second - order phase transitions . dashed ( red )
lines denote the first - order phase transitions .
shaded ( blank ) area denotes the insulator ( metal ) state .
the inset shows the result of mott scenario within the brinkman
rice criterion .
, scaledwidth=49.0% ] and @xmath0 in hfa .
notation is the same as in fig .
[ fig : mit ] . ,
scaledwidth=49.0% ] the type - iii af phase remains the ground state in a wide range @xmath111 at large @xmath16 .
this state is characterized by the presence of finite gap in electron spectrum and thus is the insulator .
a decrease of @xmath16 leads to the transition into the gapless metallic state which is magnetic at @xmath112 and paramagnetic at @xmath113 .
the most important result is that mit is of first order ( discontinuous ) , _
e.g. _ the corresponding magnetic transition demonstrates discontinuous change of magnetic state .
this differs fcc lattice from sc and bcc lattices where the second order mit is found within the same model @xcite .
a possibility of the first order mit in cubic lattices was discussed in ref .
@xcite , but concrete calculations were not performed yet .
the region of magnetic metallic state at @xmath112 consists of two regions of vertical incommensurate spiral states @xmath47 , small pocket of @xmath48 spiral state and the type - i af phase region with @xmath114 .
@xmath47 state and the type - i af phase transform into each other continuously through the second order phase transition .
a decrease of @xmath16 leads to the second order transition to the paramagnetic ( pm ) phase .
the strong increase @xmath115 changes the large-@xmath16 insulator phase from the type - iii af state to the type - ii antiferromagnet with @xmath15 . at positive @xmath56
, type - ii af phase goes to the pm metallic state with decreasing @xmath16 through the first order transition . at negative @xmath56
we have a completely different scenario : decrease of @xmath16 closes the electron spectrum gap and the type - ii af insulator smoothly goes to the type - ii af metal through the second order transition .
further on , type - ii af metal undergoes the second order transition to the @xmath44-state .
the inset in fig .
[ fig : mit ] presents the paramagnetic mit obtained from the brinkman rice criterion @xmath116 , @xmath117 being the fermi function .
therefore we state that the mott scenario is irrelevant at parameter values studied , since characteristic values of @xmath118 are much larger than those for magnetic mit .
the same conclusion is valid for the square , simple cubic and bcc lattices @xcite . for comparison we constructed an analogous mit phase diagram using hfa ( fig .
[ fig : mit_hf ] ) . in agreement with the conclusions of ref .
@xcite , the main correlation effect is the renormalization of @xmath16 , which results in a shift of magnetic regions to larger @xmath16 s within sba . for negative @xmath0
this is the only change of the hfa diagram as compared to sba . for the positive-@xmath56 half of the diagram the differences are more significant :
the hfa diagram has additional region of spiral @xmath48 state ; the metallic state is always magnetic at mit ; for @xmath119 a narrow region of type - iii af metal phase exists within hfa going into type - iii af insulator through the second order transition .
physical reason of vanishing of metallic magnetic phase region below af insulator and above paramagnetic metal regions for @xmath120 within sba is the following : from fig .
6 it is clear that metallic magnetic states are slightly more favorable within hfa than non - magnetic ones .
main shortcoming of hfa is strong overestimation of non - magnetic state energy due to artificial impossibility of reduction of the doubly occupied states number @xmath121 .
the only way to reduce the corresponding large coulomb energy within hfa is inducing the magnetization @xcite .
correct account of coulomb energy of non - magnetic states within the sba results in the suppression of magnetic metallic states .
note that the band narrowing @xmath122 for af insulator states is close to 1 and its characteristics are close to hfa ones these states are almost not influenced by correlations effects , see details below ( fig .
[ fig : aux]a ) . analogous mit phase diagrams for sc and bcc lattices @xcite demonstrate tending of @xmath54 to zero at @xmath123 . in fcc lattice
@xmath54 is always finite because this is not bipartite lattice .
mit in sc and bcc lattices is a continuous second order transition without a change of the magnetic state which is @xmath63 nel antiferromagnet ( type - ii af ) .
in contrast , mit in fcc lattice is of the first order in most of the studied @xmath56 range , except for the region of large negative @xmath56 , where we have a picture similar to other cubic lattices , _
i.e. _ continuous mit without change of magnetic order . this can be interpreted as follows : at large absolute values of @xmath56 the electron spectrum of fcc lattice tends to the the spectrum of sc lattice @xcite , and this tendency results in a similar magnetic state with @xmath15 and similar scenario of mit . according to this analogy one may expect the @xmath63 metal region at large positive values of @xmath0 , but magnetism in this region is possible only at very high @xmath16 values where existence of the gapless magnetic state is impossible .
dependence of physical quantities on both sides of the mit boundary .
solid ( dashed ) green , blue and purple lines denote @xmath124 , the magnetic moment @xmath39 and @xmath125 factor of the insulator ( metallic ) state correspondingly . vertical dotted lines separate the regions of different magnetic states .
upper ( lower ) labels denote the insulator ( metal ) magnetic states .
letters a and b denote the regions of @xmath48 and @xmath46 spiral states correspondingly .
@xmath46 here is equivalent to @xmath126 due to the symmetry of electron spectrum .
( b ) the dependence of @xmath127 vs @xmath0 for the magnetic ( solid line ) and paramagnetic ( dashed line ) scenarios of mit .
, title="fig:",scaledwidth=49.0% ] dependence of physical quantities on both sides of the mit boundary .
solid ( dashed ) green , blue and purple lines denote @xmath124 , the magnetic moment @xmath39 and @xmath125 factor of the insulator ( metallic ) state correspondingly . vertical dotted lines separate the regions of different magnetic states .
upper ( lower ) labels denote the insulator ( metal ) magnetic states .
letters a and b denote the regions of @xmath48 and @xmath46 spiral states correspondingly .
@xmath46 here is equivalent to @xmath126 due to the symmetry of electron spectrum .
( b ) the dependence of @xmath127 vs @xmath0 for the magnetic ( solid line ) and paramagnetic ( dashed line ) scenarios of mit . ,
title="fig:",scaledwidth=49.0% ] fig .
[ fig : aux]a illustrates the @xmath0 dependences of the metal and insulator characteristics at the mit boundary .
the jump of magnetic moment at mit from a magnetic to the paramagnetic phase ( @xmath128 ) is particularly large and reaches 0.9 in @xmath129 units .
if mit occurs between two magnetic states the jump is also appreciable and mostly varies from 0.2 to 0.4 . for the insulator state
the @xmath122-factor is close to 1 in the whole range of @xmath0 which implies weakness of correlation effects .
this agrees with the conclusion of paper @xcite about small role of correlation effects in vicinity of half - filling .
at the same time for metallic state @xmath122 is noticeably smaller , especially for positive @xmath56 , so that the electron correlations are more significant for @xmath130 .
[ fig : aux]b shows the critical @xmath54 coulomb interaction parameter in units of bandwidth @xmath131 for the magnetic and paramagnetic ( brinkman rice ) scenarios of mit .
note that @xmath131 depends on the @xmath56 value if it has negative sign ( @xmath132 ) , while for positive @xmath56 the dependence is absent ( @xmath133 ) .
it is clearly seen that the magnetic mit occurs at approximately two times smaller values of @xmath54 than the paramagnetic mit in all the range of @xmath0 .
interesting that the shape of both curves is very similar , indicating that @xmath0-dependence of correlation effects is qualitatively the same for both magnetic and paramagnetic phases .
it should be noted that @xmath134 at @xmath135 in afm - ii phase is almost independent on @xmath0 , which implies that system goes into a different regime .
this agrees with that the phase diagram for @xmath83 ( fig .
[ fig : fcc_t=+0.3 ] ) is more strongly modified by correlations ( as compared to the @xmath92 diagram , fig .
[ fig : fcc_t=0 ] ) than @xmath84 diagram , fig .
[ fig : fcc_t=-0.3 ] .
one can observe the rapid growth of @xmath54 ( in units of @xmath131 , fig .
[ fig : aux]b ) simultaneously with considerable suppression of @xmath122-factor in metallic state ( fig .
[ fig : aux]a ) , which can be interpreted as an enhancement of correlation effects for @xmath130 .
a quantitative comparison with the real compounds ( see the introduction ) is hardly possible because of simplicity of our model which does not account for multiple bands , spin - orbit effects etc .
however , a qualitative connection with some results for fullerides can be established .
cs@xmath3c@xmath4 is af insulator at ambient conditions which becomes superconducting under pressure .
the lowest unoccupied c@xmath4 molecular orbitals are threefold degenerate , but jahn teller distortion lifts the degeneracy by pairing two electrons , only one electron remaining unpaired @xcite .
therefore the system can be roughly described by a single - band model , and we can state an agreement between the observed first - order transition from af insulator to pm superconductor @xcite with the first - order transition from af insulator to pm metal at @xmath120 in fig .
[ fig : mit ] . a qualitative analogy can be carried out with mno , where the nearest and next - nearest exchange integrals are of the same order @xcite and a similar pressure induced first - order transition from type - ii af insulator to pm metal occurs @xcite .
it should be noted that the metal - insulator transition in mno is accompanied by a structural change which does not allow a direct comparison .
while generally first - order transitions imply a possibility of magnetic phase separation , no evidences of this were found for cs@xmath3c@xmath4 @xcite .
this result is confirmed by our calculations and is not surprising because the separation on underdoped and overdoped phases is practically impossible in the presence of the gap .
to conclude , we found that the account of conduction electron degrees of freedom allows to catch some interesting physical features of magnetic states which are discarded within the localized - electron model .
the electron transfer effects lift the degeneracy of magnetic states ( _ e .
g. _ , of collinear and spiral af type - iii ordering ) which is present within the localized model . at large @xmath94 the anderson kinetic exchange mechanism establishes the equivalence of the heisenberg and half - filled hubbard models , and the transition from af - iii to af - ii occurs at @xmath136 ( fig . [ fig : mit ] ) , which is in excellent agreement with the corresponding result of ref .
@xcite , @xmath137 . however , at smaller @xmath94 the phase diagram turns out to be strongly asymmetric with respect to the sign of @xmath56 and demonstrates presence of antiferromagnetic and spiral metallic phases between paramagnetic metal and antiferromagnetic insulator regions .
these metallic phases are influenced differently by correlations effects : at @xmath138 these states appears to be unstable due to bandwidth suppression and the magnetic order vanishes for most cases ; at @xmath139 these regions are stable at the cost of substantial reduction of magnetization amplitude .
we are grateful to a.n .
ignatenko for useful discussions .
the research was carried out within the state assignment of faso of russia ( theme `` quantum '' no . 01201463332 and financial program aaaa - a16 - 116021010082 - 8 ) .
this work was supported in part by the division of physical sciences and ural branch , russian academy of sciences ( projects no . 15 - 8 - 2 - 9 , 15 - 8 - 2 - 12 ) , by the russian foundation for basic research ( projects no .
16 - 02 - 00995 , 16 - 42 - 180516 ) and act 211 government of the russian federation 02.a03.21.0006 .
the main amount of calculations was performed using the `` uran '' cluster of imm ub ras .
0 k. knpfle , l. m. sandratskii , and j. kbler 2000 _ phys .
b _ * 62 * , 5564 ; s.v .
okatov , yu.n .
gornostyrev , a.i . lichtenstein and m.i .
katsnelson 2011 _ phys . rev .
b _ * 84 * , 214422 ; a.v .
ruban , m.i .
katsnelson , w. olovsson , s.i .
simak , and i.a .
abrikosov 2005 _ phys .
b _ * 71 * , 054402 ; p.a .
igoshev , a.v .
efremov , a.i .
poteryaev , a.a .
katanin and v.i .
anisimov 2013 _ phys .
b _ * 88 * , 155120 | the ground - state magnetic phase diagram ( including collinear and spiral states ) of the single - band hubbard model for the face - centered cubic lattice and related metal - insulator transition ( mit ) are investigated within the slave - boson approach by kotliar and ruckenstein . the correlation - induced electron spectrum narrowing and a comparison with a generalized hartree - fock approximation allow one to estimate the strength of correlation effects .
this , as well as the mit scenario , depends dramatically on the ratio of the next - nearest and nearest electron hopping integrals @xmath0 .
in contrast with metallic state , possessing substantial band narrowing , insulator one is only weakly correlated .
the magnetic ( slater ) scenario of mit is found to be superior over the mott one . unlike simple and body - centered cubic lattices , mit is the first order transition ( discontinuous ) for most @xmath0 .
the insulator state is type - ii or type - iii antiferromagnet , and the metallic state is spin - spiral , collinear antiferromagnet or paramagnet depending on @xmath0 .
the picture of magnetic ordering is compared with that in the standard localized - electron ( heisenberg ) model .
_ keywords _ : hubbard model , metal - insulator transition , slave bosons , non - collinear magnetism , antiferromagnetism , frustration |
this poster centers on the development of cloudy , a large - scale code designed to compute the spectrum of gas in photo - ionization or collisional balance .
such plasma is far from equilibrium , and its conditions are set by the balance of a host of micro - physical processes .
the development of cloudy is a three - pronged effort requiring advances in the underlying atomic data base , the numerical and computational methods used in the simulation , and culminating in the application to astronomical problems .
these three steps are strongly interwoven . in recent years
, cloudy has undergone a major upgrade in the atomic data .
a large part of this effort was concentrated on improving the photo - ionization cross sections . since
then the code has been transformed from a fortran 77 to an ansi c code and work has begun to completely recode the model atoms used to calculate line emissivities .
they will be organized in such a way that each iso - electronic sequence has a common code appropriate for all members of that sequence .
this makes validating the code easier and also facilitates future upgrades in the atomic data .
work on the hydrogen sequence is complete and work on the helium sequence is underway
. a more detailed account of the current status of the code can be found in section 2 . the grain model in cloudy
is also currently undergoing a major upgrade , which is described in more detail in section 3 .
cloudy has undergone a major transformation in the past year with respect to the last release cloudy 90 .
the most important differences are listed below .
a first release of cloudy 94 is currently available from the cloudy web - site ( for details see section 4 ) . *
the code is now strict ansi / iso 89 c. as a result cloudy is now exceptionally gnu gcc and linux friendly .
after this version is released the development version will move to c++ as gcc evolves onto the ansi / iso c++ standard and the c++ standard template library matures . *
all hydrogenic species hi through znxxx , and their respective ions are treated with a common model atom that uses a single code base .
this atom reproduces accurate hydrogenic emissivity to within the uncertainties in the atomic data .
the 30 hydrogenic ions can have up to 400 levels . *
the continuum now extends down to 10@xmath0 ryd ( @xmath1 m ) .
this is needed because the continuum must extend to the energy of the 400 399 transition of hydrogen . * previous versions used simple approximations for the hydrogenic ionization and level balance for temperatures too low to compute nlte departure coefficients .
the new version determines level populations for low temperatures rather than departure coefficients , so low temperature predictions are as valid as high temperature results .
* two additional sets of nlte stellar atmospheres are available the costar grid of wind - blanketed o - stars , and the rauch low - metallicity grid of white dwarf atmospheres . *
the optimization algorithm phymir is included .
this allows optimization runs to be executed much more efficiently on parallel unix computers by calculating individual models simultaneously on different processors .
this can drastically reduce the amount of wallclock time needed , depending on the number of free parameters . *
the ionization / thermal kernel has been totally rewritten to incorporate all the lessons learned from known convergence problems . as a result c94 is more stable than c90 , with far better convergence properties . *
all previous versions only considered ionization stages that could be produced by the incident continuum .
this limit has been lifted , so collisional or coronal equilibrium models can be computed with very soft incident continua . *
much of the code is now double precision . as a result
the code will work for a broader range of densities than before .
densities well below 10@xmath2 @xmath3 or above 10@xmath4 @xmath3 can be computed without under / over flow . *
the assert command has been introduced .
this tells the code to verify that its predictions agree ( within a stated uncertainty ) with a known result .
the test cases make extensive use of this feature , which provides an automatic way to validate the code . as of now ,
the entire test suite of standard models is recomputed and verified every single night . *
all large storage items are dynamically allocated at run time , taking only the needed memory . as a result , in its default state
, c94 actually takes less memory than c90 .
it also executes slightly faster than c90 .
the current grain model was introduced to cloudy in 1990 to facilitate more accurate modeling of the orion nebula ( for a detailed description see baldwin et al . , 1991 , apj 374 , 580 ) .
the model has since undergone some minor revisions but remained largely the same .
recently , our knowledge of grains has been greatly advanced by the results from the iso mission . in view of these rapid developments
we have undertaken a major upgrade of the grain model in cloudy .
the two main aims are to make the code more flexible and to make the modeling results more realistic . in particular , in the current model
the grain opacities for a handful of grain species are hard - wired in the code .
furthermore , only a single equilibrium temperature is calculated for each species .
this is inappropriate for a grain size distribution since small grains will be hotter than large grains .
this is caused by the fact that grain opacities depend quite strongly on grain size , as shown in figure 1 . to improve the model ,
the following changes are being implemented : * we will include a spherical mie code in cloudy .
the necessary optical constants needed to run the code will be read from a separate file .
this allows greater freedom in the choice of grain species .
files with optical constants for a range of materials will be included in the cloudy distribution .
however , the user can also supply his own optical constants for a completely new grain type .
* we will introduce mixing laws to the code .
this will allow the user to define grains which are mixtures of different materials .
cloudy will then calculate the appropriate opacities by combining the optical constants of these grain types .
this will allow the user to simulate aggregate or ` fluffy ' grains . *
the absorption and scattering opacities will be calculated by cloudy for more arbitrary grain size distributions .
this will give the user considerably more freedom .
currently only a standard ism and a truncated orion size distribution are included in the code . *
the size distribution will be split up in many small bins , and an equilibrium temperature will be calculated for each bin separately .
this allows non - equilibrium heating to be treated and more realistic grain emission spectra to be calculated .
first tests show that under realistic conditions this can make large differences in the flux ( at least a factor of two ) in the wien tail of the grain emission .
the total flux emitted by the grains remains virtually unchanged however .
the source for cloudy can be obtained from the web at the following url : | this paper describes the development of the photo - ionization code cloudy since the last major release 90.05 . # |
in the standard model ( sm ) , the process @xmath6 is symmetric under the exchange of @xmath7 and @xmath8 at leading order ( lo ) .
when next - to - leading order ( nlo ) processes are included , there is a small forward - backward asymmetry ( fba ) , of @xmath9 @xcite .
this non - zero and positive asymmetry means that ( anti-)top quarks are emitted preferably in the incoming ( anti-)quark direction . in 2011 , cdf @xcite and d collaborations @xcite measured a high fba in @xmath0 production from proton - antiproton collisions . the d asymmetry is @xmath10 and the cdf asymmetry is @xmath11 .
furthermore , cdf reported a mass dependent asymmetry @xcite : @xmath12 on the other hand , the charge asymmetry measured at atlas ( @xmath13 @xcite ) and cms ( @xmath14 @xcite ) agrees well with the sm predictions .
there are various new physics ( np ) models to explain the fb asymmetry , many of which are in tension with the lhc charge asymmetry , like sign top production , and the @xmath0 cross section . in this paper
we will consider and modify one of the light axigluon models suggested by tavares and schmaltz @xcite , which is still a viable candidate @xcite .
axigluons have a long history @xcite and there has been a significant amount of work to explain the @xmath0 fba via massive color octets @xcite . for details of the model ,
see @xcite . to summarize
, the model has an extra @xmath15 symmetry group , hence the gauge symmetry is @xmath16
. introduced with this extra symmetry group is an extra set of up- and down- type quarks , and a scalar field @xmath17 , which acquires a vacuum expectation value ( _ vev _ ) to break @xmath18 into the diagonal @xmath19 of the sm . through this symmetry breaking , one combination of the two @xmath15 gauge fields acquires a mass .
this massive color octet is called the axigluon , and its massless counterpart is the sm gluon .
similarly , there are combinations of fermions that become exotic heavy quarks and the sm light quarks , allowing the axigluon coupling to the light quarks to be a free parameter . also , there are no gauge anomalies due to cancellations from the additional quarks
. gluons couple to both the sm and heavy quarks with the same strength as expected .
the lepton sector is exactly the same as the sm , and will not be mentioned throughout this paper .
the axigluon in this model can have mass below 450 gev .
however , to be viable , it needs to be broad . in @xcite ,
the authors introduce new heavy quarks and color adjoint scalars that the axigluon can decay into .
these exotic quarks and scalars then decay into multi - jets which is not ruled out by lhc searches yet .
note also that axigluons with mass @xmath20 and enhanced couplings to top quarks can be seen via lhc four - top searches @xcite .
however this axigluon is fairly light , and it does not have enhanced top couplings . in @xcite ,
the authors consider only flavor universal couplings of axigluons to the sm quarks .
this relies on the strong assumption of an underlying global symmetry .
this global symmetry is only approximate .
mixing of heavy and light quarks could induce flavor changing neutral currents ( fcncs ) .
furthermore , since the mixing occurs between quarks that have the same @xmath21 charge , it does not give rise to flavor changing @xmath22 couplings .
therefore we will not assume the existence of an exact global symmetry of the axigluon couplings , which allows flavor changing couplings of the axigluons .
the new scalars in this model do not induce fcncs , so the axigluon couplings are the most significant possible source of new fcnc .
other models that have flavor changing color - octet couplings have been proposed in the literature @xcite .
a general lagrangian with flavor violating axigluon interactions contains the following terms : @xmath23 here @xmath24 is the axigluon , @xmath25 and @xmath26 are sm up- and down - type quarks respectively ( @xmath27 is the generation index ) , and @xmath28 are flavor independent couplings .
color and spinor indices are suppressed for simplicity .
the complex matrices @xmath29 contain off diagonal axigluon couplings of up- and down - quarks respectively .
this mixing follows from flavor symmetry breaking of heavy and light quarks .
these fcncs , which can occur at tree level , can have interesting effects on fcnc observables .
the rest of the paper is organized as follows : in section 1 , we investigate the contribution of light axigluons to @xmath1 and @xmath2 fbas . in section 2 ,
we constrain the flavor changing axigluon couplings from neutral meson mixing data . in section 3
, we investigate the contribution of the constrained flavor changing axigluon model to the following decays : @xmath3 , @xmath30 , and @xmath5 .
amplitude . left diagram is from the sm gluon exchange , right diagram is the axigluon exchange . ] in the tevatron experiments cdf and d , @xmath0 production happens mostly via @xmath6 ( fig.[tta ] ) .
the square amplitude of the process @xmath6 including the axigluon contribution is calculated as follows @xcite : @xmath31 \notag\\ & \hspace{10pt } + ng_s^4\frac{\hat{s}^2(-\hat{s}+m_a^2)^2}{((-\hat{s}+m_a^2)^2+\gamma_a^2m_a^2)^2}\biggl[\bigl[g_v^{q\,2}+g_a^{q\,2}\bigr]\bigl[g_v^{t\,2}(1+c^2 + 4m^2)+g_a^{t\,2}(1+c^2 - 4m^2)\bigr]+8g_v^tg_a^tg_v^qg_a^q\,c\biggr]\end{aligned}\ ] ] where @xmath32 is the color sum , @xmath33 is the partonic total momentum , @xmath34 is the velocity of the top quark , @xmath35 where @xmath36 is the angle between the incoming @xmath37- and the outgoing @xmath7-quark , @xmath38 is the axigluon mass , and @xmath39 is the width .
vector and axial couplings of the axigluon are defined as : @xmath40 in eq.[ttax ] , the first term comes from the sm gluon exchange , the second term is the interference between the gluon and the axigluon channels , and the third term is the axigluon s - channel ( see fig.[tta ] ) .
the fba comes from the terms that are proportional to the odd powers of @xmath41 , since @xmath41 is odd under @xmath42 .
in order to accommodate the measured @xmath0 fba , the axigluon should give a large fba without affecting the @xmath0 cross - section , which is close to its sm value . a light axigluon ( @xmath43 gev ) with a large width @xmath44 ,
is shown to agree with both the @xmath0 fba and the @xmath0 cross section @xcite .
assuming approximate parity symmetry , @xmath45 needs to be small .
we will , as in @xcite , take @xmath46 .
we will define the fba , @xmath47 , through the forward and backward scattering cross - sections : @xmath48 where @xmath49 is the total cross - section .
then the fba is @xmath50 ^ 2}g_a^4\bigl(1+{\beta^2\over3}-4m^2\bigr)}\label{fba}\end{aligned}\ ] ] where @xmath51 .
this fba is plotted in fig.[tasym ] as a function of the @xmath0 invariant mass @xmath52 .
the values for coupling constants for each axigluon mass ( 100 , 200 , 300 , 400 gev ) are taken from @xcite , where the authors performed fits to both the tevatron fba and the lhc charge asymmetry , and chose the coupling constants that would best fit both of them .
forward backward asymmetry , @xmath47 vs @xmath0 invariant mass @xmath52 .
@xmath47 only changes sign when @xmath53 .
the values for the coupling constant @xmath54 are taken from @xcite . ]
an obvious prediction of the light axigluon model is fbas for bottom and charm quarks .
although these may be challenging to measure , their observation would be an important clue .
the graphs for these asymmetries can be seen in fig.[basym ] and fig.[casym ] .
since both @xmath55 gev and @xmath56 gev are very small compared to the predicted axigluon mass ( 100s gevs ) , the fba structure is almost the same for both quarks . as can be seen from fig.[nasym ] , the predicted fbas are quite large even for low energies , @xmath57 200 gev , so the search need[line - through]*s * not [ line - through]*to * go to high invariant masses . furthermore , the crossing at @xmath58 ( @xmath59 ) is expected to be clearer compared to @xmath0 production , since @xmath60 .
consequently , measuring the @xmath1 fba would be a good way to find out the axigluon mass .
measuring the @xmath1 fba is also suggested in @xcite .
in this section we will use the data available from neutral meson mixings ( @xmath61 mixing ) to constrain the fc axigluon coupling matrices , @xmath62 and @xmath63 , in eq.[fcc ] .
neutral mesons mix " because the flavor eigenstates ( @xmath64 , @xmath65 ) of the sm hamiltonian are not the actual mass eigenstates ( @xmath66 , @xmath67 and @xmath68 for light and heavy respectively ) .
the mass difference between @xmath69 and @xmath70 will be one of the constraints we will use for each mixing . since these processes are fcnc processes they happen via loop diagrams in the sm , like electroweak ( ew ) box diagrams in fig.[ewbox ] .
these diagrams are suppressed by cabibbo - kobayashi - maskawa ( ckm ) matrix elements due to flavor changing ew interactions . however , with flavor changing axigluon couplings we have neutral meson mixing at tree level . therefore the fc couplings are constrained by the mixing data for @xmath71 and @xmath72 and @xmath73 mixing . for the following calculations we take @xmath74 gev , which gives us the least stringent constraints .
we explain the methods we use to constrain each coupling constant in the respective sections . in each section
we try to give the most general forms that can be used to constrain these couplings , but at the end we assume axial couplings for simplicity . we prefer to give constraints on real and imaginary parts of @xmath75 ( @xmath76 quark flavors ) rather than @xmath77 itself , since the meson mixing amplitudes that we use for the constraints involve the square of the coupling constants .
one can find the constraints on the real and imaginary parts of the coupling constants themselves , assuming neither of them are zero .
a summary of the results can be seen in table .
[ table ] . mixing ] .origin of constraints and upper bounds on the imaginary part and the magnitude of fc axigluon couplings [ cols="^,^,^,^",options="header " , ] mixing ] the sm and the axigluon contribution to @xmath78 mixing amplitude , where @xmath79 for @xmath80 and @xmath81 respectively , can be seen in fig.[bbmixing ] .
the gray box in the first diagram is the sum of all ew box diagrams that have @xmath82-quarks in the loop .
however , for @xmath71 mixing , the @xmath7-loop is the most important one .
the sm contribution , including nlo corrections , to this amplitude is given in @xcite as : @xmath83(\bar{b}q)_{v - a}(\bar{b}q)_{v - a}\end{aligned}\ ] ] where @xmath84 is the fermi constant , @xmath85 gev is the w - boson mass , @xmath86 are the ckm matrix elements that mix @xmath87 and @xmath88type quarks , @xmath89 is the strong structure constant evaluated at @xmath90 , and @xmath91 ( @xmath92 for @xmath93 , @xmath94 being the number of active flavors at the mixing scale ) is a constant that comes from the running of the coupling coefficients .
also @xmath95 , and @xmath96 with @xmath97 .
this function @xmath98 is one of many similar functions called inami - lin functions .
they are loop functions that arise in box and penguin diagrams in the sm and were calculated by inami and lin in @xcite . for @xmath99 gev and @xmath85 gev , @xmath100 .
thus the sm mixing amplitude is : @xmath101 the second diagram in fig.[bbmixing ] is the axigluon contribution to the mixing amplitude .
this part can be written as : @xmath102\biggr\}\end{aligned}\ ] ] where @xmath103 is the number of colors , and @xmath104 , @xmath105 and @xmath106 are the renormalization group ( rg ) evolved coefficients for the corresponding 4-quark operators .
these coefficients can be calculated by following @xcite .
let us define the following quantity : @xmath107 in order to compare the new physics ( np ) and the sm contributions to the mixing process .
we need the following matrix elements @xcite : [ meb ] @xmath108 where @xmath109 is the @xmath110 meson mass , @xmath111 is the decay constant , and @xmath112 the bag parameter ( @xmath113 for b - meson decays ) .
recent lattice - calculated values for these constants can be found in @xcite and references therein .
using eq.[meb ] in eq.[smbbar2 ] and [ baxmix ] we can write the sm and axigluon matrix elements for the mixing as follows : @xmath114m_{_{b_q^0}}f_{b_q}^2 \notag \\ & \simeq \frac{m_{_{b_q^0}}f_{b_q}^2 } { m_a^2}\left(-0.18((g_{bq}^l)^2+(g_{bq}^r)^2)-0.73\,g_{bq}^lg_{bq}^r\right)\end{aligned}\ ] ] thus eq.[delta ] reads : @xmath115 as in the flavor conserving part of the light axigluon model , we assume axial couplings : @xmath116 : @xmath117 in @xcite one can find re(@xmath118 ) and im(@xmath118 ) for @xmath119 and @xmath81 mixing . in the next two subsections we are going to look at both cases separately . in this paper
we use the ckm basis given in @xcite . in this basis , the relevant ckm matrix elements , @xmath120 and @xmath121 , are real , which makes the sm contribution real .
hence , the real and imaginary parts of @xmath122 can be constrained separately from the real and imaginary parts of @xmath123 . from @xcite , we take the following boundaries : @xmath124 to be conservative , all parameters are taken at the @xmath125 boundaries of the fits from @xcite , since there is some tension with the sm at @xmath126 . using these values and eq.[deltaq2 ] with @xmath127 , we get the following constraints : @xmath128 for @xmath129 mixing , one of the relevant ckm matrix elements , @xmath130 , is complex , therefore the real and imaginary parts of @xmath131 can not be constrained separately .
let us work this through in more detail .
we can write eq.[deltaq2 ] as follows : @xmath132 where @xmath133 notice that the real and imaginary parts of @xmath134 are very similar , so we will take @xmath135 .
then the constraint equations are [ bdc ] @xmath136 again from @xcite , we take the following bounds ( at @xmath125 ) : @xmath137 note that , for the @xmath138 , the sm value of 0 is slightly outside the @xmath125 allowed region , but we choose to disregard this discrepancy in setting the limits , since it is very small . now , notice that @xmath139 , and @xmath140 .
putting these all together with eq.[bdc ] , we get similar constraints for @xmath141 and @xmath142 such that @xmath131 lies in an approximate circle of radius @xmath143 .
so we get @xmath144 mixing ] the last constraint on @xmath145 comes from @xmath146 mixing .
the sm and the axigluon contributions to @xmath146 mixing are shown in fig.[kkmixing ] .
the lo sm contribution again comes from ew box diagrams , but in this case @xmath147- and @xmath7-quark loops both need to be considered .
the nlo mixing amplitude from the sm is @xcite : @xmath148(\bar{s}d)_{v - a}(\bar{s}d)_{v - a}\end{aligned}\ ] ] where @xmath149 is another inami - lin function , @xmath150 , @xmath151 , and @xmath152 gev ) .
the relevant matrix elements for kaon mixing can be found in @xcite : [ mek ] @xmath153 using eq.[mek ] in eq.[smkbar1 ] we get @xmath154\end{aligned}\ ] ] the axigluon contribution to kaon mixing can be written using eq.[baxmix ] with the substitution of appropriate quark operators . then using eq.[mek ]
the axigluon matrix element becomes : @xmath155 assuming again axial couplings : @xmath156 , the axigluon matrix element is : @xmath157 now , following a common notation @xcite , we define @xmath158 this matrix element @xmath159 is the off - diagonal element of the mass matrix " in the full hamiltonian @xmath160 , and @xmath161 is the effective hamiltonian , that includes both the sm and the np interactions , for the mixing process .
the off - diagonal elements of @xmath162 are related to the mass difference between heavy and light mesons , and the off - diagonal elements of @xmath163 are related to the decay of these mesons .
the interested reader should refer to @xcite and references therein for more information on meson mixings and decays .
the cp violation in meson mixings comes from a possible phase difference between @xmath159 and @xmath164 , which depends only on the short distance part of the matrix element @xmath159 .
the long distance interactions , which come from on - shell states in the loops ( fig.[ewbox ] ) , are cp conserving .
the axigluon does not contribute to the long distance part of this amplitude at lo .
the mass difference , @xmath165 , between heavy and light mesons is @xmath166 notice that @xmath165 gets affected by both the short distance and the long distance parts of the effective hamiltonian .
unfortunately , calculation of the long distance contributions is difficult @xcite . in this paper
we assume that long distance contributions are at most @xmath167 of the total mass difference , hence @xmath168 . in order to constrain the coupling constant @xmath169 ,
we use the mass difference @xcite and the imaginary part of @xmath159 @xcite : @xmath170 consequently , we get the following constraints : @xmath171 finally , we can write @xmath145 constraints as follows : @xmath172 where @xmath173 is calculated assuming @xmath174 and @xmath175 . mixing ] @xmath176 mixing , like
the other meson mixings , happens via ew box diagrams in the sm , shown in fig.[ddmixing ] .
calculation of the lo sm contribution to the mixing amplitude is similar to @xmath146 mixing where @xmath177 and @xmath178loops are changed with @xmath179- and @xmath180-loops .
@xmath176 mixing suffers more from long distance contributions as compared to @xmath181 and @xmath81 mixing .
consequently , nlo corrections to the mixing amplitude are not calculated in the literature . in this section
, we only consider the lo short distance amplitude , and assume that long distance effects are at most the same order as short distance ones .
the lo short distance sm contribution to the mixing amplitude ( at @xmath182 gev ) is : @xmath183 for the matrix elements , we assume similar relations to eq.[meb ] since @xmath184 gev .
the axigluon contribution can be written using eq.[baxmix ] with the substitution of the appropriate quark operators : @xmath185 once again , we assume axial couplings @xmath186 : @xmath187 constraints on @xmath188 come from the @xmath176 mass difference , @xmath189 gev @xcite , and the ratio @xmath190 , where @xmath37 and @xmath191 are the coefficients that describe the flavor eigenstates @xmath192 in terms of mass eigenstates @xmath193 : @xmath194 one can show that @xmath195 in @xmath176 system , @xmath196 @xcite , and
so we have @xmath197 we use eq.[qp ] to constrain the imaginary part of @xmath188 .
the real part is already more constrained from the mass difference . from @xcite , we take the following values for the magnitude and the argument of @xmath190 : @xmath198 the constraints on @xmath188 from these values are as follows : @xmath199 unfortunately , there are no other mesons with which we can investigate the up - sector further . however , the neutral @xmath200-meson system has other interesting features , like the cp asymmetry in @xmath30 decays that was measured in 2011 @xcite .
we look more into the contribution of fc axigluons to this process in the next section .
in this section , we will check the effects of the fc couplings on several sm processes , namely the rare decay @xmath3 , the cp asymmetry in @xmath30 decays , and the isospin violation in @xmath201 decays .
we do not expect a significant contribution to @xmath3 decay from the axigluons , since it is affected through axigluon - penguin loops at lo .
however , when fc axigluon currents exist , processes like @xmath30 and @xmath201 can happen at tree level .
therefore one would expect to get an appreciable contribution from axigluon induced channels .
these decays are chosen because they are of current experimental interest @xcite . in the sm
, this decay is predicted to happen very rarely , with a branching ratio of @xmath202 @xcite .
this is very close the recently measured branching ratio of @xmath203 by lhcb @xcite .
these new results constrain the np one can have that would affect this branching ratio . as we will see in this section , axigluon contribution to this decay amplitude is at least 3 orders of magnitude smaller than the sm contribution .
consequently , the measured branching ratio data does not impose further constraints on the coupling constant @xmath204 .
this branching ratio is so small in the sm as it occurs through ew penguin and box diagrams .
the sm effective hamiltonian , including nlo corrections , is calculated in @xcite as : @xmath205 where @xmath206 and @xmath207 is the weinberg angle . in eq.[y1 ] , @xmath208 where @xmath209 is the renormalization scale ( @xmath210 ) . for @xmath99 gev and @xmath85 gev , @xmath211 , and eq.[bmusm1 ]
can be written as : @xmath212 ] the axigluon contribution to this decay is shown in fig.[bmuaxf ] , which is a penguin loop with an axigluon .
this is the lo contribution , since the axigluon does not couple to leptons . in this diagram ,
@xmath213 and @xmath214loops are more important than the @xmath215loop , since the @xmath215loop is doubly suppressed by the small coupling constant .
one might worry about the divergence of the corresponding loop integral .
this divergence arises because the part of the theory we are considering is not complete , for example there are additional ( heavy ) quarks in the full theory , and their inclusion should lead to cancellations of these divergences .
the full theory is renormalizable @xcite , so we do not worry about the terms that can be canceled through the short distance contributions of the heavy quarks .
hence the amplitude ( for the @xmath213 quark loop ) is : @xmath216\right]b\end{aligned}\ ] ] where @xmath217 and @xmath218 .
adding the @xmath214 loop , and realizing that @xmath219 for @xmath220 , we neglect the part @xmath221 , and write as an @xmath222 estimate @xmath223\end{aligned}\ ] ] where @xmath224 together with @xmath225 and @xmath226 . in this paper
, we take @xmath227 .
thus the axigluon contribution to the @xmath3 decay hamiltonian becomes @xmath228 now we can compare eq.[bmuax2 ] and eq.[bmusm2 ] , for @xmath74 gev ( at @xmath229 scale ) : @xmath230 the ratio gets smaller for smaller axigluon masses .
since the axigluon loop contributions are very small compared to the sm , the uncertainties in the calculations should not be a worry .
hence flavor changing axigluon couplings under already considered constraints do not affect the @xmath3 branching ratio in a noticeable way , and so this decay does not give further constraints on the coupling constant @xmath204 . ] the cp asymmetry in @xmath30 decays is defined as follows : @xmath231 where @xmath232 is the partial width of the @xmath233 decay . in 2011 ,
the lhcb measured this asymmetry to be @xmath234 . in moriond 2013 ,
the lhcb presented a new , smaller measurement @xcite .
there still appears to be much confusion about the origin of this asymmetry in the sm @xcite .
the long distance effects in @xmath200-decays make the calculation of relevant hadronic matrix elements very difficult .
furthermore , the charm quark might not be heavy enough to trust perturbation theory at this scale .
naively , one would expect the cp violation at lo to come from decay channels like the first diagram in fig.[dcp ] , which is estimated in @xcite to give @xmath235 .
the discrepancy between @xmath236 ( experiment ) and @xmath237 ( sm ) might be due to new physics .
however it also might be contained in the sm if some matrix elements of penguin operators are much larger than the estimates given by dimensional analysis @xcite . in this section
, we compare the contributions of the two diagrams in fig.[dcp ] to the cp asymmetry .
the effective hamiltonian that comes from the sm diagram in fig.[dcp ] can be written as follows @xcite : @xmath238 where @xmath239 and @xmath240 is another inami - lin function .
the subscripts of the coefficients and the choice of writing the quark operators in this way is a slight variation of what buras does in his paper @xcite .
he gathers operators and their coefficients in a way that is easier to keep track of in the rg flow equations .
here we do not rg flow the coefficients , and compare the sm and the axigluon parts at @xmath241 gev ) .
the effective axigluon hamiltonian that can be written from the diagram on the right in fig.[dcp ] is : @xmath242 where @xmath243 @xcite .
comparing eq.[cpax ] with eq.[cpsm ] for @xmath74 gev , we see that the upper bound for axigluon contribution is an order of magnitude larger than the sm contribution .
this upper bound grows with decreasing axigluon mass , to @xmath244 times the sm contribution for @xmath245 gev .
thus this could produce larger than the sm cp violation in @xmath200-meson decays . in the sm , @xmath246 decay follows , at lo , from ew penguin and box diagrams that do not involve the @xmath247-quark , which is then called the spectator quark ( fig.[nonisosm ] ) .
for this decay , an observable , the isospin asymmetry @xmath248 , can be defined as follows : @xmath249 a similar asymmetry is defined also for @xmath250 .
we can see that in the spectator quark approximation , this asymmetry is zero , since there is no difference between the decay of the neutral and charged @xmath251meson .
if we consider diagrams in which the final @xmath252 pair is emitted from the spectator quark ( fig.[iso ] ) , there would be a non - zero isospin asymmetry due to the different charges of the spectator quarks involved in neutral and charged @xmath251meson decays . in the sm
, the asymmetry for @xmath250 is expected to be around @xmath253 @xcite .
although there is no clear prediction for the isospin asymmetry in @xmath254 from the sm , one might expect it to be similarly small , almost zero @xcite .
the isospin asymmetry that is measured at the lhcb is consistent with the sm for the @xmath250 , however it deviates from zero with @xmath255 significance for @xmath254 @xcite .
the sm prediction might be enhanced by more precise hadronic matrix element calculations , however there might as well be np involved in these decays , like flavor changing axigluons ( fig.[isoax ] ) .
decays ] in order to compare the axigluon contribution to this isospin violating process with the sm contribution , we assume that the ew penguin diagram in fig.[isosm ] is as important as any other isospin violating diagram in the sm , if not the most important one .
therefore , instead of performing comprehensive calculations of the sm contributions , we only compare the two diagrams that are shown in fig.[iso ] .
furthermore , we only consider the parts of these diagrams that are responsible for @xmath256 decay , since the emission of the final muons are the same in both cases .
[ h ] the sm contribution to the hadronic part of the effective amplitude from fig.[isosm ] can be written as follows : @xmath257 where @xmath258 .
the axigluon part of the same amplitude from fig.[isoax ] is : @xmath259 comparing eq.[isosm1 ] and eq.[isoax1 ] for @xmath74 gev ( at @xmath229 scale ) , we get @xmath260 the axigluon contribution is at most the same order as the sm one when @xmath245 gev .
the light axigluon model is an experimentally allowed modification of the sm and a viable explanation of the cdf @xmath0 forward - backward asymmetry . in this paper
we used the axigluon model suggested in @xcite to predict @xmath1 and @xmath2 forward - backward asymmetries .
they are expected to be large and depend on the invariant mass of the quark pair .
this mass dependence is a useful tool to investigate the mass of the axigluon .
we also modified this flavor conserving axigluon model to include flavor violating couplings between the axigluon and the sm quarks .
these couplings are constrained by neutral meson mixings , and the upper bounds on their magnitudes are in the range @xmath261 . after taking the upper bounds for the couplings
, we checked their effects on the rare decay @xmath3 , the cp violation in @xmath30 , and the isospin violation in @xmath5 .
we found that the fc axigluon has virtually no effect on the decay @xmath3 , since this process still occurs via loop diagrams .
this result agrees with the last measurements of the branching ratio of this decay @xcite . in the case of the isospin violation in @xmath5 ,
fc axigluon effects seem to be at most the same order as the sm ones even though the axigluon contribution is at tree level .
the most interesting effect of the fc axigluon is on cp violation in @xmath30 decays . for this cp violating asymmetry ,
the upper bound on the axigluon contribution is at least 10 times larger than the @xmath235 sm prediction .
we conclude that adding small flavor violating effects to the light axigluon model might contribute to the cp violation in @xmath30 and to neutral meson mixings .
i would like to thank ann nelson for inspiring this work and for our many fruitful conversations .
i would also like to thank dan amidei for conversations about @xmath0 and @xmath1 forward - backward asymmetry measurements at the tevatron .
this work was supported in part by the u.s .
department of energy under grant no .
de - fg02 - 96er40956 .
l. g. almeida , g. f. sterman , and w. vogelsang , `` threshold resummation for the top quark charge asymmetry , '' http://dx.doi.org/10.1103/physrevd.78.014008[_phys.rev._ * d78 * ( 2008 ) 014008 ] , http://arxiv.org/abs/0805.1885[arxiv:0805.1885 [ hep - ph ] ] .
o. antunano , j. h. kuhn , and g. rodrigo , `` top quarks , axigluons and charge asymmetries at hadron colliders , '' http://dx.doi.org/10.1103/physrevd.77.014003[_phys.rev._ * d77 * ( 2008 ) 014003 ] , http://arxiv.org/abs/0709.1652[arxiv:0709.1652 [ hep - ph ] ] .
bowen , s.d .
ellis , and d. rainwater , `` standard model top quark asymmetry at the fermilab tevatron , '' http://dx.doi.org/10.1103/physrevd.73.014008[_phys.rev._ * d73 * ( 2006 ) 014008 ] , http://arxiv.org/abs/hep-ph/0509267[arxiv:hep-ph/0509267 [ hep - ph ] ] .
w. hollik and d. pagani , `` the electroweak contribution to the top quark forward - backward asymmetry at the tevatron , '' http://dx.doi.org/10.1103/physrevd.84.093003[_phys.rev._ * d84 * ( 2011 ) 093003 ] , http://arxiv.org/abs/1107.2606[arxiv:1107.2606 [ hep - ph ] ] .
collaboration , t. aaltonen _ et al .
_ , `` evidence for a mass dependent forward - backward asymmetry in top quark pair production , '' http://dx.doi.org/10.1103/physrevd.83.112003[_phys.rev._ * d83 * ( 2011 ) 112003 ] , http://arxiv.org/abs/1101.0034[arxiv:1101.0034 [ hep - ex ] ] .
http://www-cdf.fnal.gov/physics/new/top/public_tprop.html .
collaboration , v. m. abazov _
et al . _ ,
`` forward - backward asymmetry in top quark - antiquark production , '' http://dx.doi.org/10.1103/physrevd.84.112005[_phys.rev._ * d84 * ( 2011 ) 112005 ] , http://arxiv.org/abs/1107.4995[arxiv:1107.4995 [ hep - ex ] ] .
, `` measurement of the charge asymmetry in top quark pair production in pp collisions at sqrt(s ) = 7 tev using the atlas detector , '' http://dx.doi.org/10.1140/epjc/s10052-012-2039-5 [ _ arxiv:1203.4211 _ ( mar . , 2012 ) ] . http://arxiv.org/abs/1203.4211 . eur.phys.j . c72 ( 2012 )
, `` inclusive and differential measurements of the t t - bar charge asymmetry in proton - proton collisions at sqrt(s ) = 7 tev , '' http://dx.doi.org/10.1016/j.physletb.2012.09.028 [ _ arxiv:1207.0065 _ ( june , 2012 ) ] .
b 717 ( 2012 ) 129 .
j. drobnak , j. f. kamenik , and j. zupan , `` flipping t tbar asymmetries at the tevatron and the lhc , '' http://dx.doi.org/10.1103/physrevd.86.054022[_phys.rev._ * d86 * ( 2012 ) 054022 ] , http://arxiv.org/abs/1205.4721[arxiv:1205.4721 [ hep - ph ] ] .
g. marques tavares and m. schmaltz , `` explaining the t - tbar asymmetry with a light axigluon , '' http://dx.doi.org/10.1103/physrevd.84.054008 [ _ phys.rev .
_ * d84 * ( 2011 ) 054008 ] , http://arxiv.org/abs/1107.0978[arxiv:1107.0978 [ hep - ph ] ] . c. gross , g. m. tavares , c. spethmann , and m. schmaltz , `` light axigluon explanation of the tevatron ttbar asymmetry and multijet signals at the lhc , '' http://arxiv.org/abs/1209.6375[arxiv:1209.6375 [ hep - ph ] ] .
p. h. frampton , j. shu , and k. wang , `` axigluon as possible explanation for @xmath262 forward - backward asymmetry , '' http://dx.doi.org/10.1016/j.physletb.2009.12.043[_phys.lett._ * b683 * ( 2010 ) 294297 ] , http://arxiv.org/abs/0911.2955[arxiv:0911.2955 [ hep - ph ] ] .
h. frampton and s. l. glashow , `` chiral color : an alternative to the standard model , '' http://dx.doi.org/10.1016/0370-2693(87)90859-8[_phys.lett._ * b190 * ( 1987 ) 157 ] . j. h. kuhn and g. rodrigo , `` charge asymmetry of heavy quarks at hadron colliders , '' http://dx.doi.org/10.1103/physrevd.59.054017 [ _ phys.rev .
_ * d59 * ( 1999 ) 054017 ] , http://arxiv.org/abs/hep-ph/9807420[arxiv:hep-ph/9807420 [ hep - ph ] ] .
b. xiao , y .- k .
wang , and s .- h .
zhu , `` new color - octet vector boson ?
, '' http://arxiv.org/abs/1011.0152[arxiv:1011.0152 [ hep - ph ] ] .
r. barcelo , a. carmona , m. masip , and j. santiago , `` stealth gluons at hadron colliders , '' http://dx.doi.org/10.1016/j.physletb.2011.12.002 [ _ phys.lett . _ * b707 * ( 2012 ) 8891 ] , http://arxiv.org/abs/1106.4054[arxiv:1106.4054 [ hep - ph ] ]
. j. aguilar - saavedra and m. perez - victoria , `` shaping the top asymmetry , '' http://dx.doi.org/10.1016/j.physletb.2011.10.004[_phys.lett._
* b705 * ( 2011 ) 228234 ] , http://arxiv.org/abs/1107.2120[arxiv:1107.2120 [ hep - ph ] ] .
g. z. krnjaic , `` very light axigluons and the top asymmetry , '' http://dx.doi.org/10.1103/physrevd.85.014030[_phys.rev._ * d85 * ( 2012 ) 014030 ] , http://arxiv.org/abs/1109.0648[arxiv:1109.0648 [ hep - ph ] ] .
s. dutta , a. goyal , and m. kumar , `` top quark physics in vector color - octet model , '' http://arxiv.org/abs/1209.3636[arxiv:1209.3636 [ hep - ph ] ] .
l. j. hall and a. e. nelson , `` heavy gluons and monojets , '' http://dx.doi.org/10.1016/0370-2693(85)90487-3[_phys.lett._ * b153 * ( 1985 ) 430 ] .
e. alvarez , l. da rold , j. i. s. vietto , and a. szynkman , `` phenomenology of a light gluon resonance in top - physics at tevatron and lhc , '' http://dx.doi.org/10.1007/jhep09(2011)007[_jhep_ * 1109 * ( 2011 ) 007 ] , http://arxiv.org/abs/1107.1473[arxiv:1107.1473 [ hep - ph ] ] .
chen and g. faisel , `` axigluon on like - sign charge asymmetry @xmath263 , fcncs and cp asymmetries in b decays , '' http://dx.doi.org/10.1016/j.physletb.2011.01.010[_phys.lett._ * b696 * ( 2011 ) 487494 ] , http://arxiv.org/abs/1005.4582[arxiv:1005.4582 [ hep - ph ] ] .
u. haisch and s. westhoff , `` massive color - octet bosons : bounds on effects in top - quark pair production , '' http://dx.doi.org/10.1007/jhep08(2011)088[_jhep_ * 1108 * ( 2011 ) 088 ] , http://arxiv.org/abs/1106.0529[arxiv:1106.0529 [ hep - ph ] ] .
d. kahawala , d. krohn , and m. j. strassler , `` measuring the bottom - quark forward - central asymmetry at the lhc , '' http://dx.doi.org/10.1007/jhep01(2012)069[_jhep_ * 1201 * ( 2012 ) 069 ] , http://arxiv.org/abs/1108.3301[arxiv:1108.3301 [ hep - ph ] ] .
g. buchalla , a. j. buras , and m. e. lautenbacher , `` weak decays beyond leading logarithms , '' http://dx.doi.org/10.1103/revmodphys.68.1125 [ _ rev.mod.phys . _
* 68 * ( 1996 ) 11251144 ] , http://arxiv.org/abs/hep-ph/9512380[arxiv:hep-ph/9512380 [ hep - ph ] ] .
t. inami and c. lim , `` effects of superheavy quarks and leptons in low - energy weak processes @xmath264 and @xmath265 , '' http://dx.doi.org/10.1143/ptp.65.297[_prog.theor.phys._ * 65 * ( 1981 ) 297 ] .
j. laiho , e. lunghi , and r. s. van de water , `` lattice qcd inputs to the ckm unitarity triangle analysis , '' http://dx.doi.org/10.1103/physrevd.81.034503[_phys.rev._ * d81 * ( 2010 ) 034503 ] , http://arxiv.org/abs/0910.2928[arxiv:0910.2928 [ hep - ph ] ] .
http://latticeaverages.org/. a. j. buras , m. misiak , and j. urban , `` two loop qcd anomalous dimensions of flavor changing four quark operators within and beyond the standard model , '' http://dx.doi.org/10.1016/s0550-3213(00)00437-5[_nucl.phys._ * b586 * ( 2000 ) 397426 ] , http://arxiv.org/abs/hep-ph/0005183[arxiv:hep-ph/0005183 [ hep - ph ] ] .
a. lenz , u. nierste , j. charles , s. descotes - genon , h. lacker , _
_ , `` constraints on new physics in @xmath266 mixing in the light of recent lhcb data , '' http://dx.doi.org/10.1103/physrevd.86.033008 [ _ phys.rev . _ * d86 * ( 2012 ) 033008 ] , http://arxiv.org/abs/1203.0238[arxiv:1203.0238 [ hep - ph ] ] .
collaboration , j. beringer _
_ , `` review of particle physics ( rpp ) , '' http://dx.doi.org/10.1103/physrevd.86.010001[_phys.rev._ * d86 * ( 2012 ) 010001 ] .
f. mescia and j. virto , `` natural susy and kaon mixing in view of recent results from lattice qcd , '' http://arxiv.org/abs/1208.0534[arxiv:1208.0534 [ hep - ph ] ] .
collaboration , r. aaij _ et al .
_ , `` evidence for cp violation in time - integrated @xmath267 decay rates , '' http://dx.doi.org/10.1103/physrevlett.108.129903 , 10.1103/physrevlett.108.111602[_phys.rev.lett . _
* 108 * ( 2012 ) 111602 ] , http://arxiv.org/abs/1112.0938[arxiv:1112.0938 [ hep - ex ] ] .
collaboration , r. aaij _ et al .
_ , `` strong constraints on the rare decays @xmath268 and @xmath269 , '' http://dx.doi.org/10.1103/physrevlett.108.231801[_phys.rev.lett._ * 108 * ( 2012 ) 231801 ] , http://arxiv.org/abs/1203.4493[arxiv:1203.4493 [ hep - ex ] ] .
http://kds.kek.jp/conferencedisplay.py?confid=10807 .
collaboration , r. aaij _ et al .
_ , `` measurement of the isospin asymmetry in @xmath270 decays , '' http://dx.doi.org/10.1007/jhep07(2012)133[_jhep_ * 1207 * ( 2012 ) 133 ] , http://arxiv.org/abs/1205.3422[arxiv:1205.3422 [ hep - ex ] ] .
r. fleischer , `` on branching ratios of @xmath271 decays and the search for new physics in @xmath272 , '' http://arxiv.org/abs/1208.2843[arxiv:1208.2843 [ hep - ph ] ] .
m. golden and b. grinstein , `` enhanced cp violations in hadronic charm decays , '' http://dx.doi.org/10.1016/0370-2693(89)90353-5[_phys.lett._ * b222 * ( 1989 ) 501 ] .
cheng , `` direct cp violation in @xmath273 decays , '' http://arxiv.org/abs/1206.4369[arxiv:1206.4369 [ hep - ph ] ] .
y. grossman , a. l. kagan , and y. nir , `` new physics and cp violation in singly cabibbo suppressed d decays , '' http://dx.doi.org/10.1103/physrevd.75.036008[_phys.rev._ * d75 * ( 2007 ) 036008 ] , http://arxiv.org/abs/hep-ph/0609178[arxiv:hep-ph/0609178 [ hep - ph ] ] . a. j. buras , `` weak hamiltonian , cp violation and rare decays , '' http://arxiv.org/abs/hep-ph/9806471[arxiv:hep-ph/9806471 [ hep - ph ] ] .
a. l. kagan and m. neubert , `` isospin breaking in @xmath274 decays , '' http://dx.doi.org/10.1016/s0370-2693(02)02100-7[_phys.lett._ * b539 * ( 2002 ) 227234 ] , http://arxiv.org/abs/hep-ph/0110078[arxiv:hep-ph/0110078 [ hep - ph ] ] .
t. feldmann and j. matias , `` forward backward and isospin asymmetry for @xmath275 decay in the standard model and in supersymmetry , '' _ jhep _ * 0301 * ( 2003 ) 074 , http://arxiv.org/abs/hep-ph/0212158[arxiv:hep-ph/0212158 [ hep - ph ] ] .
m. gresham , j. shelton , and k. m. zurek , `` open windows for a light axigluon explanation of the top forward - backward asymmetry , '' http://arxiv.org/abs/1212.1718[arxiv:1212.1718 [ hep - ph ] ] .
d. atwood , `` searching for the origin of cp violation in cabibbo suppressed d - meson decays . '' | the light axigluon model is a viable candidate to explain the tevatron @xmath0 forward - backward asymmetry . in this paper
we present the forward - backward asymmetries for @xmath1 and @xmath2 systems predicted by a broad light axigluon with mass 100 - 400 gev .
furthermore , we modify this flavor universal axigluon model to include flavor changing couplings of axigluons with the sm quarks .
we constrain these couplings from the available neutral meson mixing data , and investigate their effects on the rare decay @xmath3 , cp violating @xmath4 and isospin violating @xmath5 decays .
we show that a light axigluon can contribute to the observed cp violation in @xmath4 . |
manipulation of single atoms and self - assembly techniques are very useful methods in quantum engineering .
self - assembly is determined by inter atomic interactions , which at the crystal surfaces can be direct or indirect ( i.e. induced by the substrate ) .
recent studies on adatom arrangement and their dynamics on metallic surfaces show that they experience an indirect electron - gas - mediated interactions .
electronic surface states are a source of a long range interactions , which decay with the inter
particle distance @xmath1 as @xmath2 and often have an oscillatory character @xcite .
such interactions between adatoms lead to their self - alignment in rows @xcite or in hexagonal structures @xcite .
this ordering mechanism is a good candidate to be used for constructing and manipulating nanostructure systems .
linear arrangements have unique magnetic and/or electronic properties .
the ordering dynamics and a stability of an ordered structure depend on diffusion of adatoms on the surface while diffusion in a system of adatoms is controlled by interactions between them .
the collective or chemical diffusion coefficient of adsorbed species characterizes a relaxation of the local density fluctuations in a many particle system .
it involves jumps of individual atoms ( adsorbed particles ) from one binding site to another .
theoretical description of the collective diffusion process is a complicated many body problem and various approaches have been applied to it , ranging from analytic ones based on master , fokker
planck , or kramers equations to numerical monte carlo or molecular dynamics simulations .
an important background is provided in the works of reed and ehrlich@xcite , an early summary by gomer@xcite , and in reviews by danani _
et al._@xcite and ala - nissila _
et al._@xcite .
the variational approach to the collective diffusion problem , used here , was developed in a series of works@xcite and was shown to be a very efficient tool to analyze collective diffusion problems for various types of inter particle interactions for homogeneous or inhomogeneous substrates either in one or two dimensions .
we have shown in ref . that the long range repulsive interactions can be responsible for rapid macroscopic rearrangements of adatoms upon minuscule adsorbate density changes .
we have shown that the adsorbate density dependent diffusion coefficient has peaks at densities corresponding to any ordered phase in a devil staircase phase diagram .
it is well known@xcite that the devil staircase structure emerges when the inter
particle interaction potential is repulsive and decays faster than @xmath3 .
it has been shown , one the other hand , that adatoms on metallic surfaces often interact via forces with oscillating in @xmath1 potential energies @xcite .
they exhibit a @xmath2 decay modified by friedel oscillations of the electron gas correlation function meaning that particles attract each other at some distances and repel at others .
such a distance dependence of the inter particle interactions should , in principle , result in a diffusion character similar to that already investigated for pure @xmath2 repulsion . on the other hand ,
the oscillating interaction potential has local minima .
similarly , a minimum is present also for the interaction of the lennard jones type .
the question arises how the presence of such a minimum ( or minima ) affects the diffusion kinetics in a many particle system . in particular
, it is an interesting question how does the oscillating character of the interactions decaying like @xmath2 modify diffusion in comparison with the interactions decaying like @xmath2 monotonically .
it is well known that the shape of interaction potential , in particular , existence of its attractive parts is very important for the static behavior of the system .
systems with attractive interactions form stable clusters of ordered phases , whereas ordering via repulsion is always global it affects the entire system .
consequently , an ordered phase due to attractive interactions occurs for a wider range of adatom densities so it is easier to observe experimentally@xcite .
such difference in static properties has to affect the dynamic behavior too , so it should affect also the diffusion process . for example , a fast collective diffusion in ordered structures@xcite leads to a fast reorganization of the adsorbed layer . in this work
we compare three types of interactions : monotonically decaying like @xmath2 , the oscillating ones decaying like @xmath2 with electron - gas - mediated oscillations , and the @xmath0 lennard jones interactions .
we analyze the influence of the shape of the interaction potential on the diffusion coefficient at different adsorbate densities .
magnitude of the diffusion coefficient depends on several parameters , most of them related to the character and strength of the inter
particle interaction . in
what follows we analyze how periodic in @xmath1 variation of the potential superimposed on the @xmath2 decay influences diffusion .
we show that for systems of particles interacting via oscillating electron - gas - mediated forces or via forces corresponding to the @xmath0 lennard
jones potentials the diffusion process depends sensitively on the ratio between the distance @xmath1 at which the potential has a minimum and the substrate lattice constant .
this ratio allows to distinguish between the commensurate and incommensurate type of diffusion kinetics .
we show that the diffusion coefficient depends periodically on the characteristic length of the potentials under investigation .
the paper is organized as follows . in section
ii the approach to the diffusion coefficient calculation is shortly described .
section iii contains description of results for electron - gas mediated , oscillation potential , and then for @xmath0 lennard jones potential .
section iv summarizes main results of the work .
system of @xmath4 particles interacting via long range forces is distributed homogeneously over a one - dimensional substrate of length @xmath5 with a lattice constant @xmath6 .
the interaction of two particles at the lattice positions @xmath7 and @xmath8 contributes the potential energy @xmath9 to the total energy of the system .
following ref .
we consider systems with the pair potential energy @xmath10 decreasing rapidly with @xmath1 .
this justifies to neglecting the next - nearest - particle interactions and accounting only for pair interactions between neighbors no matter how large the intra - pair separation @xmath11 is .
the total interaction energy of the system is @xmath12 where @xmath13 is the number of nearest neighbors pairs of length @xmath14 ( in units of @xmath6 ) and only @xmath14 s satisfying the condition @xmath15 are admitted in the sum . in a grand canonical ensemble approach we let @xmath16 to vary from @xmath17 to @xmath18 and keep , instead , the system under fixed external pressure @xmath19 ( in 1d it is just an external force ) which is determined by the condition that the mean nearest neighbor pair length @xmath20 is equal to the inverse of the actual coverage @xmath21 . in such case
a probability of a pair of a length @xmath14 is@xcite @xmath22 where @xmath23 and @xmath24 is a single nearest neighbor - pair isothermal
isobaric partition function .
( [ eq : w10 ] ) through ( [ eq : w12 ] ) allow to determine the thermodynamic properties of the system .
in particular , the equation of state , relation between the coverage , pressure and temperature is obtained by evaluating the mean nearest neighbor pair length @xmath25 and identifying it with @xmath26 . in the low temperature limit
the main contributions to this sum come from one or at most two terms only@xcite .
collective diffusion of the system is modeled by a kinetic lattice gas with the particle hopping rates depending on the actual potential energy of the particle .
the potential energy landscape is build by the static potential due to the substrate , as experienced by a single particle , and by interactions of the particle with its neighbors .
time evolution of the system is controlled by a set of master rate equations for the probabilities @xmath27 that a _ microstate _ @xmath28 of a lattice gas occurs at time @xmath29 @xmath30 . \nonumber\end{aligned}\ ] ] the microstate @xmath28 is understood as a set of variables specifying which particular sites in the lattice are occupied and which are not .
@xmath31 is a transition probability per unit time ( transition rate ) that the microstate @xmath28 changes into @xmath32 due to a jump of a particle from an occupied site to an unoccupied neighboring site .
microstates @xmath28 and @xmath32 differ here only by the position of a single particle , the one which jumped .
for thermally activated jumps the hopping rate depends on the difference between the energy of the system when the hopping particle is at an intermediate position between the sites engaged in the jump and the energy of the system when the particle is in its initial position .
the only contributions that do not cancel out in the difference is the energy of the hopping particle in its initial position and its energy in the activated state at the top of the potential barrier which it jumps over .
for the particle hopping from the adsorption site specified by a pair of integers @xmath33 ( i.e. with the nearest neighbors of adsorbed particles being at a distance @xmath34 and @xmath35 , respectively , to its left and right ) to a neighboring site @xmath36 the potential energy at the initial adsorption site is @xmath37 where @xmath38 is static potential energy at given site due to the interactions with the substrate . the hopping rate can be written as @xmath39},\end{aligned}\ ] ] where @xmath40 $ ] is a hopping rate for an isolated ( i.e. non interacting ) particle and @xmath41 is an intrinsic attempt frequency .
@xmath42 is the amount by which the potential energy @xmath43 of the hopping particle at a bridge site between its initial and the final position is modified by interactions with the neighbors at each its side .
we parametrize a microstate @xmath28 as @xmath44 $ ] by selecting one particle as a _ reference particle _ , denoting its lattice position as @xmath45 and specifying positions @xmath46 of all remaining @xmath47 particles with respect to it@xcite .
@xmath48 is referred to as a _ configuration_. master equations ( [ eq:1 ] ) are linear in set of probabilities @xmath49 and so their lattice fourier transform with respect to @xmath45 can be easily done .
the result is that @xmath50components of the probabilities , @xmath51 , evolve in time independently of each other .
the rate equations for @xmath51 can be expressed in terms of a @xmath50space microscopic rate matrix @xmath52 ( with rows and columns labelled by @xmath48 ) containing the individual hopping rates and phase factors like @xmath53 .
details can be found in refs . .
the collective diffusion coefficient is related to that eigenvalue @xmath54 ( termed the diffusive eigenvalue ) of @xmath52 which vanishes like @xmath55 in the limit @xmath56 .
this eigenvalue is then estimated from above in a spirit of a variational principle@xcite as @xmath57 \cdot { \bm \phi } } { { \tilde { \bm \phi } } \cdot { \bm \phi } } \rightarrow -d k^2,\end{aligned}\ ] ] where the @xmath58 stands for the @xmath56 limit and @xmath59 and @xmath60 are variational trial left and right , respectively , eigenvectors of @xmath52 corresponding to the diffusive eigenvalue .
it has been shown @xcite that for a homogeneous substrate the @xmath48-th component of the trial left eigenvector has the form @xmath61 and that @xmath62 , where @xmath63 is the probability of the configuration @xmath48 in equilibrium .
we calculate the diffusion coefficient @xmath64 as a ratio ( the @xmath56 limit is implied ) @xmath65 of the `` expectation value '' numerator @xmath66 to the `` normalization '' denominator @xmath67 detailed balance condition was used to derive eq .
( [ eq:21 ] ) so each @xmath68 term in
it accounts for transitions between @xmath48 and @xmath69 in either direction .
therefore , each configuration pair @xmath68 appears in the sum in eq .
( [ eq:21 ] ) only once [ as indicated by the comment `` no rep '' above the sum ] in order to avoid double counting . in the grand canonical ensemble approach , mentioned earlier , both @xmath70 and @xmath71 are functions of @xmath72 , and @xmath4 .
we note in passing that @xmath70 and @xmath71 are directly related to the diffusion coefficient static ( or thermodynamic ) and kinetic factor , respectively , which the diffusion coefficient is customarily factorized into@xcite .
the former is controlled only by the static interactions , determining the equilibrium properties of the system , while the latter is also sensitive to the dynamic interactions within the adsorbate and the dynamic interactions with the substrate , both controlling the rate of an approach to the thermodynamic equilibrium .
certain characteristic features of the density dependence of the static factor , often being signatures of an onset of an organization within the system , may or may not be compensated by the features present in the kinetic factor , resulting in the density dependent diffusion coefficient from which such features may be absent .
this issue for long range repulsive inter
particle interaction was examined in detail in ref . .
it was shown@xcitethat , for the one dimensional system with long range interactions , the denominator ( [ eq:22 ] ) can be expressed as @xcite @xmath73 while the numerator can be written as @xmath74 the `` no rep '' restriction in eq .
( [ eq:21 ] ) results in only the rates of jumps from the left to right to be explicitly present in eq .
( [ eq : w28 ] ) [ alternatively , expression mathematically equivalent to ( [ eq : w28 ] ) with only the right to left jump rates explicitly appearing in it can be used ] . to evaluate @xmath75 using eq .
( [ eq : w28 ] ) the potential energy correction due to interactions of the activated particle is needed .
one of the simplest models accounting for the activated particle interactions is obtained by realizing that the particle hopping from the adsorption site @xmath33 to @xmath76 surmounts a potential energy barrier at a bridge site situated , approximately , at a distance @xmath77 and @xmath78 from its nearest left and right adsorbed particle neighbor , respectively , and by evaluating the interaction potential energy at the bridge by using @xmath79 generalized to half - integer arguments .
consequently , the potential energy correction due to interactions of a particle at the bridge site between @xmath33 and @xmath76 site is @xmath80 which , used in eq .
( [ eq : w48 ] ) , leads to the following hopping rate for the left to right jumps @xmath81}.\end{aligned}\ ] ] using ( [ eq : w45 ] ) in ( [ eq : w28 ] ) yields the kinetic factor @xmath82 ^ 2 } % \nonumber \\
\left [ \sum_{l=1}^\infty e^{-\beta \tilde{\varepsilon}(l+\frac{1}{2})}\right]^2 , \end{aligned}\ ] ] in which the definition of @xmath83 in eq .
( [ eq : w11 ] ) is used . similarly like in @xmath84 in eq .
( [ eq : w12 ] ) the main contribution to the sum over @xmath14 in eq .
( [ eq : w46 ] ) comes from one or at most two terms for low enough temperatures but , for a given value of @xmath19 , these terms may correspond to different @xmath14 s than those most significant in @xmath85 . in general , the sum has to be evaluated numerically .
it has been shown that the potential energy of interactions of adatoms adsorbed on metallic surfaces like cu / cu(111 ) @xcite , co or co on cu(111 ) @xcite , fe or co on ag(111 ) @xcite , fe / cu(111 ) @xcite and ce / ag(111 ) @xcite vary with the inter
atomic distance @xmath1 as @xmath86 where @xmath87 and @xmath88 are , respectively , the fermi energy and the fermi wave vector of the surface electrons .
shape of the potential energy depends also on the fermi
level phase shift @xmath89 which for many systems is equal to @xmath90 .
this value will be used further in our calculations .
particle distance dependence of the interaction energy ( [ potential ] ) is plotted in fig .
[ fig:1 ] using a dashed line and compared with a long range purely repulsive potential energy @xmath91 used in refs . .
the value of the parameter @xmath92 used in fig .
[ fig:1 ] has been chosen in such a way that the repulsive potential energy curve forms an upper envelope of electron gas mediated potential energy ( [ potential ] ) .
strength of both types of interactions decays with the distance like @xmath2 , however , whereas the forces corresponding to the @xmath93 potential energy are repulsive at any inter atomic separation , the potential energy ( [ potential ] ) oscillates , generating attractive forces at some inter particle distances and the repulsive ones at others .
attractive forces in the system lead to the creation of stable structures at the surface and should affect the dynamic properties of thesystem .
we compare here the diffusion kinetics in adsorbates with both types of inter particle interactions and , in addition , in systems in which the interactions correspond to the @xmath0 lennard jones potential energy @xmath94,\end{aligned}\ ] ] typical for interactions between neutral atoms .
the parameters @xmath95 and @xmath96 decide about depth and position of the single potential minimum .
the lennard
jones potential has a repulsive @xmath97 wall at short distances , much steeper than @xmath2 , as it can be seen in the fig .
[ fig:1 ] . at larger distances
this potential is attractive and , decaying like @xmath98 , it is weak in comparison with any of the other two .
it will be shown , however , that the existence of this attractive part is sufficient for the diffusion kinetics in the system with the lennard jones interactions to be qualitatively similar at some densities to that in systems with oscillating interactions .
it is known@xcite that a system of particles with long - range unscreened repulsive interactions orders at @xmath99 at densities ( coverages ) given by a rational fraction ( smaller than @xmath100 ) when the interaction potential decays faster than @xmath3 .
the coverage plotted against external potential has a fractal form called a devil staircase @xcite .
it is also known@xcite that systems with interactions decaying faster than @xmath3 but additionally screened to nearest neighbors no matter how far they are ( as described in sec .
[ sec : b ] ) also form ordered @xmath99 phases at coverages equal to @xmath101 where @xmath102 is any natural number .
the `` phase diagram '' ( density vs. pressure or the chemical potential ) in this case is not as complicated as that for the unscreened interactions it has no fractal structure .
for such a system it is possible to calculate all static properties@xcite as well as to investigate fully the collective diffusion kinetics@xcite for many models of microscopic kinetics . in general , the collective diffusion coefficient peaks at sufficiently low temperatures at densities at which the system orders ( except when sharp drops in compressibility are compensated fully by kinetics for very special models of microscopic kinetics@xcite ) . for these densities and for densities around them
( at finite temperatures ) the diffusion is very fast , meaning that particles have an ability to rearrange quickly when the density changes even by a small amount .
( color online ) inter particle interaction potential energies as a function of the inter particle distance @xmath1 .
thick solid line ( red online ) : @xmath103 with @xmath104 ; dashed line ( green online ) : the oscillating potential energy in eq .
( [ potential ] ) with @xmath105 , @xmath106 and @xmath107 ; thin solid line ( blue online ) : the @xmath0 lennard jones potential energy in eq .
( [ eq : lj ] ) with @xmath108 and @xmath109 .
energies are measured in arbitrary energy units .
, width=226 ] when the interactions are purely repulsive , like for @xmath103 , the rearrangement into an ordered phase occurs for the entire lattice gas at once , the phase transition is of the second order , and formation of ordered domains surrounded by disordered regions with lower or higher densities is not possible .
such possibility opens for systems with interactions like in eq .
( [ potential ] ) which , depending on inter particle distances alternate between repulsion and attraction . when particles attract each other at some distances it is possible to order the system locally even if globally the density is too low for that@xcite . in fig .
[ fig:2 ] we compare the coverage dependence of the diffusion coefficient plotted at two different temperatures for the interaction potential energies plotted in fig . [ fig:1 ] ( with the same parameters as there ) .
it can be seen that the character of the curves corresponding to the interaction potential energies @xmath110 and the oscillating one [ given in eq .
( [ potential ] ) ] is similar at the same temperatures .
the only significant qualitative difference between these two cases is seen around @xmath111 at the lower of the two temperatures considered : the diffusion coefficient for the oscillating interaction exhibits a peak almost invisible for the system with a purely repulsive interaction at the same temperature despite the fact that for purely repulsive interaction the system does order at @xmath112 at @xmath99 .
it indicates that the presence of the inter
particle attraction for the system with oscillating interaction allows for an ordering corresponding to @xmath112 already at temperatures much higher than those are needed for such ordering with the repulsive long range interactions only .
in fact , @xmath113 is the preferred inter particle distance ( based solely on on the total interaction energy considerations ) for the potential energy parameters selected in this example , so local ordered domains are preferentially formed which correspond to a local coverage @xmath112 .
attraction , creating the potential energy minimum , is a necessary condition for the formation of the local ordered domains .
high diffusion coefficient within such domains aids in their formation because collective diffusion of particles effectively controls the system ability to create ordered phases@xcite . with purely repulsive interactions
the ordering at @xmath112 is possible only globally so the diffusion coefficient peak , very sharp at @xmath99 , is easily smeared out by thermal fluctuations
. we can generalize our conclusions here : structures build by attractive inter
particle interactions are capable of creating local domains of ordered phases .
value of the ratio of the potential energy spatial oscillation period to the substrate lattice constant determines which of the diffusion coefficient peaks ( present for @xmath99 at @xmath114 for purely repulsive long range interactions ) is amplified or attenuated .
( color online ) coverage dependence of the collective diffusion coefficient at temperatures @xmath115 ( higher curves ) and @xmath116 ( lower curves ) for the interaction potential energies plotted in fig .
[ fig:1 ] .
thick continuous lines ( red online ) : for @xmath103 ; dashed line ( green online ) : for the oscillating potential energy in eq .
( [ potential ] ) ; thin solid line ( blue online ) : for the @xmath0 lennard jones potential energy in eq .
( [ eq : lj ] ) .
the interaction potential energy parameters are the same as in fig .
[ fig:1 ] and @xmath117 is expressed in the same arbitrary energy units as the interaction potential energy.,width=226 ] analyzing fig .
[ fig:2 ] a bit more in detail for the case of lower temperatures we see that at coverages below @xmath118 both the oscillating and the strictly repulsive interactions result in the diffusion coefficient of roughly the same magnitude except , however , around @xmath112 where the broad peak for the oscillating interactions results in a somewhat faster diffusion for this case .
both types of interactions lead to a sharp increase of the diffusion coefficient at @xmath119 another coverage of a @xmath99 ordered phase@xcite . for @xmath120
the diffusion is more efficient for the purely repulsive interactions than it is for the oscillating ones .
this can be traced back to the fact that the purely repulsive interaction has much stronger repulsive core at @xmath121 than the oscillating interaction has ( c.f .
[ fig:1 ] ) .
all these features are largely washed out by thermal fluctuations at the higher temperature of the two considered here .
we show also in fig .
[ fig:2 ] the diffusion coefficient for the lennard jones interaction . here , the interactions favor the inter
particle distance of about @xmath122 , almost the same as the oscillating interaction does so it is not surprising that at low concentrations both interactions result in almost the same value of the diffusion coefficient . at higher concentrations , however , the diffusion kinetics is controlled by a very steep repulsive core at short distances , much steeper for the lennard
jones than for any of the remaining two interactions .
consequently , the diffusion coefficient increases to very high values already at @xmath112 .
more features unique to the lennard jones interaction will be discussed in sec .
[ sec : e ] .
( color online ) inter particle potential energies ( [ potential ] ) as a function of the inter
particle interaction distance @xmath1 for several values of @xmath88 ( from @xmath123 up to @xmath124 ) and @xmath125 ( in the same arbitrary energy units as in the preceding figures.),width=264 ] ( color online ) density dependence of the diffusion coefficient in the system with oscillating inter particle interaction ( [ potential ] ) for several values of @xmath88 ( from @xmath123 up to @xmath124 the same as in fig .
[ fig:4a ] ) , @xmath125 and @xmath126 ( the latter two parameters are expressed in the same arbitrary energy units as in the preceding figures.),width=226 ] in general , the character of the density dependence of the collective diffusion coefficient strongly depends on how closely the minima and the maxima of the oscillating interaction potential energy match the distances between particles which occupy the substrate lattice sites . in fig .
[ fig:4a ] the interaction potential energies for several values of @xmath88 are shown .
it can be seen that the position of the first minimum moves towards higher inter particle distances with decreasing @xmath88 .
geometrically , the minimum inter
particle distance possible is @xmath6 , the substrate lattice constant , and the overall character of the coverage dependence of the diffusion coefficient is to a major extent determined by the character of the interaction at this particular inter
particle distance .
if at the separation @xmath6 the inter particle interaction is strongly repulsive , as it is for @xmath127 or @xmath128 , then the diffusion coefficient behaves as a function of coverage ( for coverages between @xmath129 and @xmath100 ) similarly to that corresponding to a purely repulsive interaction ( as seen already in fig .
[ fig:2 ] for @xmath130 ) : it raises rapidly when the coverage approaches the value @xmath119 from below , suffers a kink and then it decreases slowly with further increase of @xmath131 .
this behavior was already analyzed for a purely repulsive long range interaction in fig .
9 of ref . .
it can be traced back to the behavior of the diffusion coefficient static factor , proportional to @xmath132 ( i.e. proportional to an inverse of an isothermal compressibility ) , modified by the behavior of the kinetic factor , proportional to @xmath71 .
for these values of @xmath88 the particles avoid occupying the nearest neighbor sites so for @xmath129 they preferentially occupy every second lattice site .
a substantial additional pressure is needed to compress the system above half occupation of the lattice , the isothermal compressibility is very low , i.e. the static factor goes through a sharp maximum at @xmath119 .
the overall shape of the density dependence of the diffusion coefficient is determined , however , by both the static and the kinetic factor . with the hopping rates given in eq .
( [ eq : w45 ] ) the high value for interaction potential energy at the inter particle distance @xmath6 results in a kinetic factor which also increases sharply at @xmath119 but does not reach a maximum there .
instead , for @xmath133 variations of the kinetic factor almost perfectly compensate for the variations of the static factor for these densities : the kinetic factor continues to increase ( albeit , less dramatically than at @xmath119 ) , goes through a broad plateau and eventually decreases sharply as @xmath131 approaches the full coverage value of @xmath100 .
the compensation between both factors for @xmath134 results in the diffusion coefficient which only slowly decreases as a function of @xmath131 over this coverage interval .
the structure observed in fig .
[ fig:4 ] at coverages below @xmath135 ( i.e. at @xmath136 and smaller ) for @xmath137 may be understood in a similar way . in these cases
the inter
particle interaction is strongly repulsive not only at a separation @xmath6 but also at the next possible one , @xmath138 , so at coverages close to @xmath136 the particles preferentially occupy every third site with the resulting compressibility minimum at that coverage .
we can see in fig .
[ fig:4 ] that locally , within the coverage interval @xmath139 , the behavior of the diffusion coefficient for the interactions corresponding to @xmath137 is quite reminiscent of that within the interval around @xmath140 and above it ( for the same values of @xmath88 and also @xmath141 ) : a sharp increase at @xmath112 is followed by a slow diffusion coefficient decrease with increasing @xmath131 ( until the singularity at @xmath119 takes over ) . for still lower @xmath88 s similar structures are observed in fig .
[ fig:4 ] around @xmath142 and even @xmath143 but due to their overlap they are not very well resolved at the temperature selected for the plots .
comparing in fig .
[ fig:2 ] the diffusion vs. coverage curves evaluated at the same temperatures for the @xmath103 interaction with the ones for the corresponding to it ( as defined in fig .
[ fig:1 ] ) @xmath130 oscillating interaction ( [ potential ] ) we see that the presence of inter particle attraction for the latter results in more pronounced structures in @xmath144 .
the diffusion maximum around @xmath112 , very broad and almost unnoticeable for the purely repulsive interaction , becomes quite obvious and sharp for the corresponding to it oscillating interaction .
the @xmath112 diffusion coefficient peak is clearly visible in fig .
[ fig:4 ] also for @xmath145 and @xmath123 .
the diffusion coefficient increase at some adlayer densities ( here @xmath112 ) is due to the attraction felt by particles at specific distances .
the overall shape , however , of the coverage dependence of diffusion curves is always a result of several different , often compensating each other factors , like a character of interaction at different distances ( which affect both the static and the kinetic factor ) , a relative height of the interaction potential energy barrier which a hopping particle must overcome between adsorption sites ( which enters the kinetic factor ) .
it is not possible , therefore , to guess from the interaction potential energy curves alone whether maxima of the diffusion coefficient at certain coverages do exist or do not . only analyzing the diffusion coefficient vs. density curves we may rationalize the existence of certain characteristic features .
note that the structure observed around @xmath112 disappears when @xmath88 increases from @xmath146 to @xmath128 . upon further increase of @xmath88 also the structure around @xmath119
apparently , the repulsion at the inter particle distance of @xmath138 and @xmath6 , respectively , ceases to be a factor controlling behavior of diffusion .
indeed , we see in fig . [ fig:4a ] that upon increasing @xmath88 above @xmath130 the repulsion at an inter particle distance of @xmath138 weakens and is gone for @xmath147 consequently any structure around @xmath112 is no longer present .
then , with further increase of @xmath88 above @xmath147 the repulsion at a separation of @xmath6 also weakens and , consequently , no structure around @xmath119 is observed for @xmath148 .
it was checked that upon increasing @xmath88 gradually up from @xmath146 the structure around @xmath112 in fig .
[ fig:4 ] evolves initially into the locally concave dependence around @xmath112 and then into the locally convex one , as seen in fig .
[ fig:4 ] for @xmath147
. further increase of @xmath88 washes out the only remaining structure around @xmath119 , @xmath144 becomes a structureless concave function ( as seen for @xmath148 ) and then it evolves into a convex one ( as seen for @xmath149 ) .
it is worth noting that the evolution of the @xmath144 from a function with the structure at @xmath119 for @xmath147 through the concave structureless function for @xmath148 and then to the convex one for @xmath149 parallels the evolution of @xmath144 observed in figs . 3 and 4 of ref . for the short range interaction
changing from the strongly repulsive , through the weakly repulsive , to the attractive one . indeed , as seen in fig .
[ fig:4a ] , the inter particle interaction at a distance @xmath6 is strongly repulsive for @xmath147 , weakly so for @xmath148 and somewhat attractive for @xmath149 .
convex @xmath144 is characteristic for systems with dominant attractive inter
particle interactions which cause them to bond and to form clusters .
when particles stay on average at distances that minimize the total interaction energy , the jump rates are lowered according to ( [ eq : w48 ] ) resulting in a decrease of the diffusion coefficient kinetic factor .
diffusion for @xmath149 oscillating interaction decreases generally as a function of density .
the decrease , quite fast at low coverages , slows down at higher ones .
we have argued above that a sudden change of slope of @xmath144 observed in fig .
[ fig:4 ] at certain characteristic coverages can be understood as due to strong inter
particle repulsion at very short distances .
this , however , can not be the only reason for such a behavior because , except at @xmath119 , such sudden slope changes are much less pronounced for purely repulsive interactions , certainly the attraction must also play a role .
we note also that all @xmath144 curves in fig .
[ fig:4 ] for @xmath150 start as convex functions of coverage already at @xmath151 and the characteristic _ smallest _ coverage at which they suffer a sudden decrease of its slope is @xmath152 , and @xmath153 for @xmath154 , and @xmath100 , respectively . in each case , the corresponding value of @xmath155 is equal to the inter particle distance at which the potential energy ( [ potential ] ) with the appropriate value of @xmath88 has the first deepest minimum in fig .
[ fig:4a ] .
the convex character of @xmath144 at coverages @xmath156 is a signature of the attractive character of the inter particle interactions at distances beyond the minimum , i.e. the distances larger than @xmath155 . at @xmath157
the system tends to order with inter
particle distances being , on average , equal to @xmath155 . for @xmath158 the average inter
particle distances are small enough for the repulsion being the dominant interaction and ordering is still possible at some coverages resulting in @xmath144 going through the second maximum at @xmath112 ( for @xmath159 ) and @xmath119 ( for @xmath160 ) . in short , for a particular oscillating interaction ( i.e.particular @xmath88 ) , the smallest coverage @xmath161 at which @xmath144 ceases to be a convex function is a boundary between the coverages ( @xmath156 ) for which the inter particle attraction dominates and the ones ( @xmath162 ) for which the repulsion does dominate . for @xmath163
the interaction is effectively equivalent to a short range attraction ( the first minimum of the interaction potential energy is at @xmath164 ) while for @xmath148 the repulsion at @xmath121 competes with attraction at @xmath165 . for this intermediate case
the character of @xmath144 resembles , as observed earlier , the one appropriate for the weak repulsive interactions .
character of the inter
particle interaction at closest possible separations , i.e. repulsion , attraction , or the potential energy minimum , determines the overall shape of the @xmath144 dependence . for small @xmath166 s ,
the particles that reside at closest possible separations from their neighbors experience the interaction induced repulsion . with increasing @xmath166 they find themselves first at the interaction potential energy minimum , then experience attraction and then the interaction energy maximum .
the cycle repeats with further increase of @xmath166 .
such a cycle of successive repulsions and attractions at a given separation should lead to a non monotonic dependence of the diffusion coefficient on the parameter @xmath88 at a particular fixed coverage : we expect that after a monotonic @xmath166 dependence of @xmath64 for @xmath166 up to about @xmath100 the diffusion coefficient should pass through a series of minima and maxima as @xmath166 further increases .
indeed , this is observed in fig .
[ fig:3 ] in which @xmath64 is plotted against @xmath166 for fixed coverage . following the curve for @xmath112
we observe a monotonic decrease of @xmath64 until a minimum is reached for @xmath167 . with further increase of @xmath166 the maximum
is reached for @xmath168 , followed by a broad minimum around @xmath169 , a maximum around @xmath170 , and a minimum again , at @xmath171 .
the oscillations of @xmath64 continue with increasing @xmath166 but their amplitude decreases mirroring the decreasing amplitude of oscillations of the interaction potential energy @xmath172 with increasing @xmath166 [ c.f .
( [ potential ] ) ] .
for @xmath119 the @xmath88 dependence of @xmath64 is very similar to that for @xmath112 except that the former starts with much higher value for small @xmath173 , has an inflection point around @xmath168 rather than a narrow minimum followed by a narrow maximum . beyond @xmath174
both curves follow closely each other .
( color online ) diffusion coefficient @xmath64 for @xmath126 as a function of a parameter @xmath88 of the oscillating interaction potential energy ( [ potential ] ) with @xmath106 for coverages @xmath112 ( dashed line ) and @xmath119 ( solid line ) .
@xmath87 and @xmath175 are expressed in the same arbitrary energy units as in preceding figures.,width=226 ] neutral particles are known to interact via long - range lennard jones potential energy ( [ eq : lj ] ) .
its repulsive core falls off more rapidly than @xmath2 ( in fact , it falls as @xmath176 , reaches an equlibrium position minimum at @xmath177 , and is attractive at larger distances , falling off like @xmath98 ) .
a staircase of ordered phases at coverages @xmath114 is supported by such an interaction ( similarly like for a purely repulsive interaction ) , the most prominent ones being @xmath119 and @xmath178 , and due to huge repulsion at short distances the diffusion coefficient of particles ordered in such phases is abnormally large in comparison to that when the system is not ordered .
we plot in fig .
[ fig:6 ] the coverage dependence of the diffusion coefficient at fixed temperature in a system with lennard jones interactions corresponding to several values of @xmath96 ( the other parameter in eq .
( [ eq : lj ] ) , @xmath95 is not varied ) . the overall shape of the presented curves is determined by an interplay between two length parameters in the system : the distance @xmath177 below which particles repel each other , and the substrate lattice permitted minimum separation @xmath6 between interacting particles .
( color online ) coverage dependence of diffusion coefficient for the lennard
particle interaction ( [ eq : lj ] ) for @xmath116 , @xmath109 and several values of @xmath96 . @xmath95 and @xmath175 are expressed the same but arbitrary energy units.,width=226 ] we can systematically analyze shapes of the @xmath144 curves in fig .
[ fig:6 ] . for @xmath179
there are no inter particle interactions ( except for site blocking ) and @xmath180 as seen in fig .
[ fig:6 ] .
for @xmath181 we have @xmath182 so when @xmath183 the interaction between particles separated by @xmath6 is attractive . consequently , it is energetically preferable for particles to be as close to each other as possible and for such values of @xmath96 the @xmath144 dependence is convex ( as seen for @xmath184 and @xmath185 in fig .
[ fig:6 ] ) , typical for systems for which the inter particle attraction dominates ( c.f .
fig . 4 in ref . ) . for @xmath186
the interaction at the closest separation @xmath121 becomes repulsive and , indeed , for @xmath187 , the @xmath144 dependence becomes concave resembling qualitatively @xmath144 for weakly repulsive short range interaction in fig .
3 of ref . .
with further increase of @xmath96 the repulsion at @xmath121 increases : for @xmath188 we note a characteristic maximum of @xmath64 around @xmath119 also observed already for stronger short range interactions in fig . 3 of ref . .
for @xmath189 the repulsion at @xmath121 is strong enough to induce a preferential occupation of every second site ( note that the interaction at the @xmath165 separation is still weakly attractive in this case ) causing the diffusion coefficient to suddenly raise by many orders of magnitude as @xmath131 approaches @xmath153 from below . with further increase of @xmath96
we reach , at @xmath190 , the point beyond which the interaction at @xmath165 is no longer attractive . for @xmath96
somewhat larger than @xmath191 we have a very strong repulsion at @xmath121 and a much weaker one at @xmath165 .
we see in fig .
[ fig:6 ] that for @xmath192 a local maximum of @xmath144 develops for @xmath136 hinting at a preferential occupation of every third site around this coverage due to repulsion for @xmath165 , followed by a sharp , almost discontinuous raise of @xmath64 at @xmath119 due to an extremely strong repulsion at @xmath121 . with further increase of @xmath96
the repulsion at @xmath165 becomes stronger so the structure around @xmath112 becomes as sharp as that around @xmath119 .
the examples are curves for @xmath193 and @xmath194 in fig .
[ fig:6 ] . for the latter , however , the interaction at the separation @xmath113 is also repulsive .
in fact , with further increase of @xmath96 features similar to those around @xmath119 and @xmath178 develop also around @xmath195 , as seen for @xmath196 .
this is because for @xmath197 the interaction at the separation @xmath113 becomes repulsive and , when strong enough , it leads to a preferential occupation of every fourth site for @xmath198 .
this structure is not present yet for @xmath199 at the temperature selected in fig .
[ fig:6 ] .
the character of the coverage dependence of the diffusion coefficient changes qualitatively every time when the interaction between particles becomes repulsive at any of the inter particle separations permitted by the substrate lattice : i.e. every time when @xmath96 increases through @xmath200 .
for @xmath96 such that @xmath201 the particles separated by distances @xmath202 repel each other ( with the repulsion being stronger for shorter separations ) while they attract each other for separations @xmath203 and larger . in such case one
expects at @xmath204 a sharp increase of the diffusion coefficient at coverages @xmath205 due to the low temperature structural organization of the adsorbate at these coverages to minimize the total interaction energy in the system . in our runs
done at lowest temperatures for which the calculations are feasible we see such structures for @xmath119 , @xmath178 and @xmath206 .
still , one expects that as the parameter @xmath96 is varied one should observe a non monotonic oscillatory changes of the diffusion coefficient at low enough coverages .
indeed , this is seen in fig .
[ fig:5 ] in which @xmath207 is plotted as a function of @xmath208 .
the maxima are noted for @xmath208 approximately _ half way _ between @xmath209 for @xmath210 ,
i.e. for @xmath211 , for which the interaction is already strongly repulsive for all inter particle separation up to @xmath212 , respectively .
( color online ) dependence of the diffusion coefficient at @xmath213 on the lennard
jones interaction parameter @xmath96 for @xmath116 and @xmath109 ( same as in fig .
[ fig:6]).,width=226 ]
variational approach has been applied to examine collective diffusion in a one dimensional lattice gas systems with two type of long range inter
particle interaction : the electron gas mediated interaction described by the oscillating friedel like potential energy ( [ potential ] ) , and the lennard
jones interaction corresponding to the potential energy ( [ eq : lj ] ) .
we have discussed the features of the coverage ( adsorbate density ) dependence of the diffusion coefficient for both these interactions and compared them with those investigated in detail for purely repulsive long
range interaction@xcite corresponding to the potential energy @xmath214 as well as with the behavior typical for short range repulsive and attractive interactions@xcite .
in general , at densities above half coverage ( @xmath133 ) , at which the inter particle repulsion at short distances plays a main role , the diffusion coefficient for the repulsive and the oscillating interaction behaves , as a function of coverage , qualitatively similarly : the diffusion coefficient is much higher than without interactions and depends on coverage relatively weakly .
this is true even for the lennard
jones interaction , except that a very steep repulsive core in this case , making creation of a high density adsorbate energetically very costly , results in huge values of the diffusion coefficient at such densities . at lower densities ,
the behavior of the coverage dependence of diffusion coefficient for the lennard jones interaction , repulsive for inter particle separations @xmath215 and attractive for @xmath216 , is quite easy to understand . at sufficiently low temperatures ,
it experiences for increasing @xmath131 a finite series of progressively sharper increases at `` critical '' coverages @xmath114 ( @xmath217 ) .
the first one , for @xmath218 corresponds to the largest inter
particle distance @xmath219 at which the interaction between the particles is still repulsive ( i.e. at the separation @xmath220 the interaction is already attractive ) . as temperature increases ,
the structures at lower densities of this series are usually smoothed out .
in fact , we observe for the lennard
jones interaction a delicate interplay between two length scales : the lattice constant @xmath6 which determines what actual distances between particles are possible , and the characteristic length of the interaction potential , @xmath221 , separating the short range repulsion from the attraction at larger distances .
when @xmath222 then the repulsive core of the interaction is irrelevant and the diffusion coefficient exhibits features similar to those observed for short range attractive interactions .
features of diffusion for the oscillating interaction are somewhat more difficult to explain in simple terms due to the existence of multiple interaction potential energy minima and , consequently , alternating regions of inter particle attraction and repulsion .
still , a sudden raise of the diffusion coefficient at @xmath119 has origin similar to that for such structure for the purely repulsive and lennard
jones interaction .
the structures for @xmath223 are , however , masked by the influence which attraction alternating with repulsive interactions has on diffusion .
we must note also that interpreting features of the diffusion coefficient in terms of the features of the inter
particle interaction is deficient in that respect that it necessarily is limited to the interpretation of the kinetic phenomenon ( diffusion ) in terms of the static properties of the system , i.e. in terms of the static ( or thermodynamic ) factor commonly used in theories of diffusion .
the kinetic factor is known , however , to be capable of compensating for very often drastic behavior of the static factor as a function of density @xcite .
this is also the case here : a slow variation of the diffusion coefficient as a function of coverage for several curves in figs .
[ fig:4 ] and [ fig:6 ] for @xmath131 immediately larger than @xmath224 , and @xmath206 is the result of a delicate compensation between a sharp drop of the static factor ( which at these particular coverages suffers a sharp cusp like maximum corresponding to a sharp drop in an isothermal compressibility ) and a strong increase of the kinetic factor when the coverage is increased through these values .
this work was supported by poland s ministry of science and higher education grant no .
n202 042 32/1171 .
the authors thank dr .
z. w. gortel from the university of alberta for continuing interest in the subject , discussions , and help in preparation of this manuscript .
99 e.wahlstrm,i.ekval,h.olin , and l. wallden , appl .
a * 66 * , s1107 ( 1998 ) .
z. zhang , q. niu , and c - k shih , phys .
80 * , 5381 ( 1998 ) .
p. hyldgaard and m. persson , j. phys .
c * 12 * , l13 ( 2000 ) .
j. repp et .
al . , phys . rev . lett.,*85 * , 2981 ( 2000 ) .
n. knorr et .
al . , phys .
b , * 65 * , 115420 ( 2002 ) . y. tiwary and k. a. fichthorn , phys . rev .
b * 75 * , 235451 ( 2007 ) . t. yokoyama et .
al . , phys .
, * 98 * , 206102 ( 2007 ) . a. schiffrin et .
al . , phys .
b , * 78 * , 035424 ( 2008 ) h.f ding et . al . , phys . rev .
b * 76 * , 033409 ( 2007 ) . e. bussmann et .
al . , phys .
bf 101 , 266101 ( 2008 ) n. n. negulyaev et .
al . , phys .
, * 77 * , 085430 ( 2008 )
. j. lagoute , x. liu and s. flsch , * 74 * , 125410 ( 2006 ) . c. liu , t. uchihashi , and t. nakayama , phys .
, * 101 * , 146104 ( 2008 ) .
i. fernandez - torrente et . al . , phys .
lett . * 99 * , 176103 ( 2007 ) .
f. silly et .
al . , phys .
, * 92 * , 016101 ( 2004 ) . n. n. negulyaev et .
al . , phys . rev .
* 74 * , 035421 ( 2006 ) .
d. a. reed and g. ehrlich , surf .
sci . * 102 * , 588 ( 1981 ) .
r. gomer , rep .
. phys . * 53 * , 917 ( 1990 ) .
a. danani , r. ferrando , e. scalas , and m. torri , int . j. mod
b * 11 * , 2217 ( 1997 ) .
nissila , r. ferrando , and s. c. ying , adv . phys . * 51 * , 949 ( 2002 ) .
z. w. gortel and m. a. zauska kotur , phys .
b * 70 * , 125431 ( 2004 ) .
m. a. zauska kotur and z. w. gortel , phys .
b * 72 * , 235425 ( 2005 ) .
l. badowski , m. a. zauska kotur , and z. w. gortel , phys .
b * 72 * , 245413 ( 2005 ) .
m. a. zauska kotur , .
badowski , and z. w. gortel , physica a * 357 * , 305 ( 2005 )
. m. a. zauska kotur and z. w. gortel , phys .
b * 76 * , 245401 ( 2007 ) . f. krzyewski and m. a. zauska - kotur , phys . rev .
b , * 78 * , 235406 ( 2008 ) .
m. yakes , m. hupalo , m. a. zauska - kotur , z. w. gortel , and m. c. tringides , phys .
letters * 98 * , 135504 ( 2007 ) .
m. a. zauska kotur and z. w. gortel , phys .
b * 74 * , 045405 ( 2006 )
. v.v . slavin and a. a. slutskin , phys .
b , * 54 * , 8095 ( 1996 ) .
p. bak and r. bruinsma , phys .
, * 49 * , 249 ( 1982 ) . | collective diffusion coefficient in a one dimensional lattice gas adsorbate is calculated using variational approach .
particles interact via either a long - range , or a long range electron - gas - mediated ( for a metallic substrate ) , or a @xmath0 lennard - jones interaction .
diffusion coefficient as a function of the adsorbate density strongly depends on the relationship between the substrate lattice constant and the characteristic length of the inter
particle interaction potential ( which determines positions of the potential energy minima ) .
the diffusion coefficient at fixed density as a function of the interaction characteristic length has an oscillating character due to the interplay between the inter
particle distances allowed by the substrate lattice structure and the average inter particle distances which minimize the total interaction energy . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.